{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<!--TABLE OF CONTENTS-->\n",
    "Contents:\n",
    "- [Data Preprocessing](#Data-Preprocessing)\n",
    "  - [Data scaling](#Data-scaling)\n",
    "  - [Categorical variables](#Categorical-variables)\n",
    "    - [One-hot encoding](#One-hot-encoding)\n",
    "      - [Method 1](#Method-1)\n",
    "      - [Method 2](#Method-2)\n",
    "    - [Ordinal encoding](#Ordinal-encoding)\n",
    "    - [Target Encoding](#Target-Encoding)\n",
    "  - [Multiple column transformations](#Multiple-column-transformations)\n",
    "- [Cross validation and Hyperparameter tuning](#Cross-validation-and-Hyperparameter-tuning)\n",
    "  - [Grid search](#Grid-search)\n",
    "  - [Randomized search](#Randomized-search)\n",
    "  - [Successive Halving search](#Successive-Halving-search)\n",
    "  - [Other libraries for hyperparameter tuning](#Other-libraries-for-hyperparameter-tuning)\n",
    "  - [Practice question](#Practice-question)\n",
    "- [Note](#Note)\n",
    "- [Algorithm chains and pipelines](#Algorithm-chains-and-pipelines)\n",
    "  - [Building Pipelines](#Building-Pipelines)\n",
    "  - [The General Pipeline Interface](#The-General-Pipeline-Interface)\n",
    "  - [Pipeline and ColumnTransformer](#Pipeline-and-ColumnTransformer)\n",
    "  - [Grid-Searching Which Model To Use](#Grid-Searching-Which-Model-To-Use)\n",
    "- [Splitting data in cross-validation](#Splitting-data-in-cross-validation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following exercise is adapted from Chapter 4 of the book *Introduction to Machine Learning with Python*, by Andreas C. Müller, Sarah Guido."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Regardless of the types of features our data consists of, how we represent them can have an enormous effect on the performance of machine learning models. The question of how to represent our data best for a particular application is known as feature engineering, and it is one of the main tasks of data scientists trying to solve real-world problems. Representing the data in the right way can have a bigger influence on the performance of a supervised model than the exact parameters we choose."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from collections import Counter\n",
    "from sklearn.inspection import DecisionBoundaryDisplay"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data scaling\n",
    "Some algorithms are very sensitive to the scaling of the data. \n",
    "Therefore, a common practice is to adjust the features so that the data representation is more suitable for these algorithms.\n",
    "Let's visualize two different scaling techniques: normalization and standardization.\n",
    "First, let's generate a small synthetic dataset with sklearn's *make_blobs()*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'our dataset')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.datasets import make_blobs\n",
    "X, y = make_blobs(n_samples=50, centers=2, random_state=1, cluster_std=3)\n",
    "plt.figure(figsize=(4,4))\n",
    "plt.scatter(X[:, 0], X[:, 1], s=30, c=y) \n",
    "plt.title(\"our dataset\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first feature (the x-axis value) is between -15 and 3. The second feature (the y-axis value) is between -10 and 10.\n",
    "Now, let's scale the data with two different scalers *MinMax* (performs normalization) and *StandardScaler* (performs standardization)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler, StandardScaler\n",
    "\n",
    "minmax_scaler = MinMaxScaler()\n",
    "standard_scaler = StandardScaler()\n",
    "X_minmax = minmax_scaler.fit_transform(X)\n",
    "X_standard = standard_scaler.fit_transform(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's see the difference:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, 2, figsize=(13,4))\n",
    "ax[0].set_title(\"Standard scaler\")\n",
    "ax[0].scatter(X_standard[:, 0], X_standard[:, 1], s=30, c=y)\n",
    "ax[0].axis([-2, 2, -2.5, 2.5])\n",
    "\n",
    "ax[1].set_title(\"Minmax scaler\")\n",
    "ax[1].scatter(X_minmax[:, 0], X_minmax[:, 1], s=30, c=y )\n",
    "ax[1].axis([-2, 2, -2.5, 2.5]);\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The *StandardScaler* in *sklearn* ensures that for each feature the mean is 0 and the variance is 1, bringing all features to the same magnitude. However, this scaling does not ensure any particular minimum and maximum values for the features. \n",
    "\n",
    "The *MinMaxScaler*, on the other hand, shifts the data such that all features are exactly between 0 and 1. For the two-dimensional dataset this means all of the data is contained within the rectangle created by the x-axis between 0 and 1 and the y-axis between 0 and 1."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Previously, we have mentioned that it is important that we should fit the scaler only on the training data and transform the test data. This way we avoid any data leakage.\n",
    "Now, let's also demonstrate why should we apply the same transformation to the training set and the test set for the supervised model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X, y = make_blobs(n_samples=50, centers=5, random_state=4, cluster_std=2)\n",
    "X_train, X_test = train_test_split(X, random_state=5, test_size=.1)\n",
    "\n",
    "fig, ax = plt.subplots(1, 3, figsize=(13, 4))\n",
    "ax[0].scatter(X_train[:, 0], X_train[:, 1], label=\"Training set\", s=30, c = 'blue')\n",
    "ax[0].scatter(X_test[:, 0], X_test[:, 1], marker='^', label=\"Test set\", s=30,c = 'red')\n",
    "ax[0].legend(loc='upper left')\n",
    "ax[0].set_title(\"Original Data\")\n",
    "\n",
    "# scale the data using MinMaxScaler\n",
    "scaler = MinMaxScaler()\n",
    "scaler.fit(X_train)\n",
    "X_train_scaled = scaler.transform(X_train)\n",
    "X_test_scaled = scaler.transform(X_test)\n",
    "\n",
    "# visualize the properly scaled data\n",
    "ax[1].scatter(X_train_scaled[:, 0], X_train_scaled[:, 1], label=\"Training set\", s=30, c = 'blue')\n",
    "ax[1].scatter(X_test_scaled[:, 0], X_test_scaled[:, 1], marker='^', label=\"Test set\", s=30,c = 'red')\n",
    "ax[1].legend(loc='upper left')\n",
    "ax[1].set_title(\"Scaled Data\")\n",
    "\n",
    "# rescale the test set separately\n",
    "# so test set min is 0 and test set max is 1\n",
    "# DO NOT DO THIS! For illustration purposes only.\n",
    "test_scaler = MinMaxScaler()\n",
    "test_scaler.fit(X_test)\n",
    "X_test_scaled_badly = test_scaler.transform(X_test)\n",
    "\n",
    "# visualize wrongly scaled data\n",
    "ax[2].scatter(X_train_scaled[:, 0], X_train_scaled[:, 1], label=\"training set\", s=30, c = 'blue')\n",
    "ax[2].scatter(X_test_scaled_badly[:, 0], X_test_scaled_badly[:, 1], marker='^',  label=\"test set\", s=30,c = 'red')\n",
    "ax[2].legend(loc='upper left')\n",
    "ax[2].set_title(\"Improperly Scaled Data\")\n",
    "\n",
    "for axi in ax:\n",
    "    axi.set_xlabel(\"Feature 0\")\n",
    "    axi.set_ylabel(\"Feature 1\")\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first subplot is an unscaled two-dimensional dataset, with the training set shown as blue circles and the test set shown as red triangles. The second subplot is the same data, but scaled using the *MinMaxScaler*. Here, we called `fit` on the training set, and then called `transform` on the training and test sets. We can see that the dataset in the second subplot looks identical to the first; only the ticks on the axes have changed. Now all the features are between 0 and 1. We can also see that the minimum and maximum feature values for the test data (the triangles) are not 0 and 1.\n",
    "\n",
    "The third subplot shows what would happen if we scaled the training set and test set separately. In this case, the minimum and maximum feature values for both the training and the test set are 0 and 1. But now the dataset looks different. The test points moved, as they were scaled differently. We changed the arrangement of the data in an arbitrary way. Clearly this is not what we want to do. Hence, **we should use `fit` on the training set only, and then `transform` on the training and test sets**."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Categorical variables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For this part of the exercise, we will use the dataset of adult incomes in the United States, derived from the 1994 census database. The task of the dataset is to predict whether a worker has an income of over 50,000 or under 50,000.\n",
    "Place the downloaded csv file *census_data* in the same directory as your notebook and import it.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>capital-gain-category</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>occupation</th>\n",
       "      <th>income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>cat2</td>\n",
       "      <td>40</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>cat1</td>\n",
       "      <td>13</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>cat1</td>\n",
       "      <td>40</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>cat1</td>\n",
       "      <td>40</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>cat1</td>\n",
       "      <td>40</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age         workclass  education capital-gain-category  hours-per-week  \\\n",
       "0   39         State-gov  Bachelors                  cat2              40   \n",
       "1   50  Self-emp-not-inc  Bachelors                  cat1              13   \n",
       "2   38           Private    HS-grad                  cat1              40   \n",
       "3   53           Private       11th                  cat1              40   \n",
       "4   28           Private  Bachelors                  cat1              40   \n",
       "\n",
       "          occupation income  \n",
       "0       Adm-clerical  <=50K  \n",
       "1    Exec-managerial  <=50K  \n",
       "2  Handlers-cleaners  <=50K  \n",
       "3  Handlers-cleaners  <=50K  \n",
       "4     Prof-specialty  <=50K  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('census_data.csv')\n",
    "# For illustration purposes, we will only select some of the columns\n",
    "data = data[['age', 'workclass', 'education',  'capital-gain-category', 'hours-per-week', 'occupation', 'income']]\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32561, 7)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age                       int64\n",
       "workclass                object\n",
       "education                object\n",
       "capital-gain-category    object\n",
       "hours-per-week            int64\n",
       "occupation               object\n",
       "income                   object\n",
       "dtype: object"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The *workclass*, *education*, *\tcapital-gain-category*, and *occupation* features are categorical features. All of them come from a fixed list of possible values, as opposed to a range, and denote a qualitative property, as opposed to a quantity. In our example, target variable is income, and it is also categorical."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One-hot encoding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By far the most common way to represent categorical variables is using the one-hot-encoding or one-out-of-N encoding, also known as dummy variables. The idea behind dummy variables is to replace a categorical variable with one or more new features that can have the values 0 and 1. We can represent any number of categories by introducing one new feature per category, as described here. In general, this encoding works well if there are only a few levels.\n",
    "For illustration purposes, we will do one hot encoding for all categorical features, even though education is ordinal, as the values it can take can be ordered."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's see the values that the *workclass* feature can take."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "workclass\n",
       "Private             22696\n",
       "Self-emp-not-inc     2541\n",
       "Local-gov            2093\n",
       "?                    1836\n",
       "State-gov            1298\n",
       "Self-emp-inc         1116\n",
       "Federal-gov           960\n",
       "Without-pay            14\n",
       "Never-worked            7\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.workclass.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's apply pandas function *get_dummies* to do one hot encoding."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32561, 7)\n",
      "(32561, 48)\n",
      "Index(['age', 'hours-per-week', 'workclass_?', 'workclass_Federal-gov',\n",
      "       'workclass_Local-gov', 'workclass_Never-worked', 'workclass_Private',\n",
      "       'workclass_Self-emp-inc', 'workclass_Self-emp-not-inc',\n",
      "       'workclass_State-gov', 'workclass_Without-pay', 'education_10th',\n",
      "       'education_11th', 'education_12th', 'education_1st-4th',\n",
      "       'education_5th-6th', 'education_7th-8th', 'education_9th',\n",
      "       'education_Assoc-acdm', 'education_Assoc-voc', 'education_Bachelors',\n",
      "       'education_Doctorate', 'education_HS-grad', 'education_Masters',\n",
      "       'education_Preschool', 'education_Prof-school',\n",
      "       'education_Some-college', 'capital-gain-category_cat1',\n",
      "       'capital-gain-category_cat2', 'capital-gain-category_cat3',\n",
      "       'capital-gain-category_cat4', 'occupation_?', 'occupation_Adm-clerical',\n",
      "       'occupation_Armed-Forces', 'occupation_Craft-repair',\n",
      "       'occupation_Exec-managerial', 'occupation_Farming-fishing',\n",
      "       'occupation_Handlers-cleaners', 'occupation_Machine-op-inspct',\n",
      "       'occupation_Other-service', 'occupation_Priv-house-serv',\n",
      "       'occupation_Prof-specialty', 'occupation_Protective-serv',\n",
      "       'occupation_Sales', 'occupation_Tech-support',\n",
      "       'occupation_Transport-moving', 'income_<=50K', 'income_>50K'],\n",
      "      dtype='object')\n",
      "age                             int64\n",
      "hours-per-week                  int64\n",
      "workclass_?                      bool\n",
      "workclass_Federal-gov            bool\n",
      "workclass_Local-gov              bool\n",
      "workclass_Never-worked           bool\n",
      "workclass_Private                bool\n",
      "workclass_Self-emp-inc           bool\n",
      "workclass_Self-emp-not-inc       bool\n",
      "workclass_State-gov              bool\n",
      "workclass_Without-pay            bool\n",
      "education_10th                   bool\n",
      "education_11th                   bool\n",
      "education_12th                   bool\n",
      "education_1st-4th                bool\n",
      "education_5th-6th                bool\n",
      "education_7th-8th                bool\n",
      "education_9th                    bool\n",
      "education_Assoc-acdm             bool\n",
      "education_Assoc-voc              bool\n",
      "education_Bachelors              bool\n",
      "education_Doctorate              bool\n",
      "education_HS-grad                bool\n",
      "education_Masters                bool\n",
      "education_Preschool              bool\n",
      "education_Prof-school            bool\n",
      "education_Some-college           bool\n",
      "capital-gain-category_cat1       bool\n",
      "capital-gain-category_cat2       bool\n",
      "capital-gain-category_cat3       bool\n",
      "capital-gain-category_cat4       bool\n",
      "occupation_?                     bool\n",
      "occupation_Adm-clerical          bool\n",
      "occupation_Armed-Forces          bool\n",
      "occupation_Craft-repair          bool\n",
      "occupation_Exec-managerial       bool\n",
      "occupation_Farming-fishing       bool\n",
      "occupation_Handlers-cleaners     bool\n",
      "occupation_Machine-op-inspct     bool\n",
      "occupation_Other-service         bool\n",
      "occupation_Priv-house-serv       bool\n",
      "occupation_Prof-specialty        bool\n",
      "occupation_Protective-serv       bool\n",
      "occupation_Sales                 bool\n",
      "occupation_Tech-support          bool\n",
      "occupation_Transport-moving      bool\n",
      "income_<=50K                     bool\n",
      "income_>50K                      bool\n",
      "dtype: object\n"
     ]
    }
   ],
   "source": [
    "data_dummies = pd.get_dummies(data)\n",
    "print(data.shape)\n",
    "print(data_dummies.shape)\n",
    "print(data_dummies.columns)\n",
    "print(data_dummies.dtypes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that the continuous features *age* and *hours-per-week* were not touched, while the categorical features were expanded into one new feature for each possible value. The *get_dummies* function in pandas treats all numbers as continuous and will not create dummy variables for them. If we have a categorical feature with values as numbers, we should then explicitely list it for dummy coding."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's expand the output display for pandas, such that all the columns are shown when we print the dataframe. By default, not all columns will be displayed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>workclass_?</th>\n",
       "      <th>workclass_Federal-gov</th>\n",
       "      <th>workclass_Local-gov</th>\n",
       "      <th>workclass_Never-worked</th>\n",
       "      <th>workclass_Private</th>\n",
       "      <th>workclass_Self-emp-inc</th>\n",
       "      <th>workclass_Self-emp-not-inc</th>\n",
       "      <th>workclass_State-gov</th>\n",
       "      <th>workclass_Without-pay</th>\n",
       "      <th>education_10th</th>\n",
       "      <th>education_11th</th>\n",
       "      <th>education_12th</th>\n",
       "      <th>education_1st-4th</th>\n",
       "      <th>education_5th-6th</th>\n",
       "      <th>education_7th-8th</th>\n",
       "      <th>education_9th</th>\n",
       "      <th>education_Assoc-acdm</th>\n",
       "      <th>education_Assoc-voc</th>\n",
       "      <th>education_Bachelors</th>\n",
       "      <th>education_Doctorate</th>\n",
       "      <th>education_HS-grad</th>\n",
       "      <th>education_Masters</th>\n",
       "      <th>education_Preschool</th>\n",
       "      <th>education_Prof-school</th>\n",
       "      <th>education_Some-college</th>\n",
       "      <th>capital-gain-category_cat1</th>\n",
       "      <th>capital-gain-category_cat2</th>\n",
       "      <th>capital-gain-category_cat3</th>\n",
       "      <th>capital-gain-category_cat4</th>\n",
       "      <th>occupation_?</th>\n",
       "      <th>occupation_Adm-clerical</th>\n",
       "      <th>occupation_Armed-Forces</th>\n",
       "      <th>occupation_Craft-repair</th>\n",
       "      <th>occupation_Exec-managerial</th>\n",
       "      <th>occupation_Farming-fishing</th>\n",
       "      <th>occupation_Handlers-cleaners</th>\n",
       "      <th>occupation_Machine-op-inspct</th>\n",
       "      <th>occupation_Other-service</th>\n",
       "      <th>occupation_Priv-house-serv</th>\n",
       "      <th>occupation_Prof-specialty</th>\n",
       "      <th>occupation_Protective-serv</th>\n",
       "      <th>occupation_Sales</th>\n",
       "      <th>occupation_Tech-support</th>\n",
       "      <th>occupation_Transport-moving</th>\n",
       "      <th>income_&lt;=50K</th>\n",
       "      <th>income_&gt;50K</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>40</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50</td>\n",
       "      <td>13</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>40</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53</td>\n",
       "      <td>40</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28</td>\n",
       "      <td>40</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  hours-per-week  workclass_?  workclass_Federal-gov  \\\n",
       "0   39              40        False                  False   \n",
       "1   50              13        False                  False   \n",
       "2   38              40        False                  False   \n",
       "3   53              40        False                  False   \n",
       "4   28              40        False                  False   \n",
       "\n",
       "   workclass_Local-gov  workclass_Never-worked  workclass_Private  \\\n",
       "0                False                   False              False   \n",
       "1                False                   False              False   \n",
       "2                False                   False               True   \n",
       "3                False                   False               True   \n",
       "4                False                   False               True   \n",
       "\n",
       "   workclass_Self-emp-inc  workclass_Self-emp-not-inc  workclass_State-gov  \\\n",
       "0                   False                       False                 True   \n",
       "1                   False                        True                False   \n",
       "2                   False                       False                False   \n",
       "3                   False                       False                False   \n",
       "4                   False                       False                False   \n",
       "\n",
       "   workclass_Without-pay  education_10th  education_11th  education_12th  \\\n",
       "0                  False           False           False           False   \n",
       "1                  False           False           False           False   \n",
       "2                  False           False           False           False   \n",
       "3                  False           False            True           False   \n",
       "4                  False           False           False           False   \n",
       "\n",
       "   education_1st-4th  education_5th-6th  education_7th-8th  education_9th  \\\n",
       "0              False              False              False          False   \n",
       "1              False              False              False          False   \n",
       "2              False              False              False          False   \n",
       "3              False              False              False          False   \n",
       "4              False              False              False          False   \n",
       "\n",
       "   education_Assoc-acdm  education_Assoc-voc  education_Bachelors  \\\n",
       "0                 False                False                 True   \n",
       "1                 False                False                 True   \n",
       "2                 False                False                False   \n",
       "3                 False                False                False   \n",
       "4                 False                False                 True   \n",
       "\n",
       "   education_Doctorate  education_HS-grad  education_Masters  \\\n",
       "0                False              False              False   \n",
       "1                False              False              False   \n",
       "2                False               True              False   \n",
       "3                False              False              False   \n",
       "4                False              False              False   \n",
       "\n",
       "   education_Preschool  education_Prof-school  education_Some-college  \\\n",
       "0                False                  False                   False   \n",
       "1                False                  False                   False   \n",
       "2                False                  False                   False   \n",
       "3                False                  False                   False   \n",
       "4                False                  False                   False   \n",
       "\n",
       "   capital-gain-category_cat1  capital-gain-category_cat2  \\\n",
       "0                       False                        True   \n",
       "1                        True                       False   \n",
       "2                        True                       False   \n",
       "3                        True                       False   \n",
       "4                        True                       False   \n",
       "\n",
       "   capital-gain-category_cat3  capital-gain-category_cat4  occupation_?  \\\n",
       "0                       False                       False         False   \n",
       "1                       False                       False         False   \n",
       "2                       False                       False         False   \n",
       "3                       False                       False         False   \n",
       "4                       False                       False         False   \n",
       "\n",
       "   occupation_Adm-clerical  occupation_Armed-Forces  occupation_Craft-repair  \\\n",
       "0                     True                    False                    False   \n",
       "1                    False                    False                    False   \n",
       "2                    False                    False                    False   \n",
       "3                    False                    False                    False   \n",
       "4                    False                    False                    False   \n",
       "\n",
       "   occupation_Exec-managerial  occupation_Farming-fishing  \\\n",
       "0                       False                       False   \n",
       "1                        True                       False   \n",
       "2                       False                       False   \n",
       "3                       False                       False   \n",
       "4                       False                       False   \n",
       "\n",
       "   occupation_Handlers-cleaners  occupation_Machine-op-inspct  \\\n",
       "0                         False                         False   \n",
       "1                         False                         False   \n",
       "2                          True                         False   \n",
       "3                          True                         False   \n",
       "4                         False                         False   \n",
       "\n",
       "   occupation_Other-service  occupation_Priv-house-serv  \\\n",
       "0                     False                       False   \n",
       "1                     False                       False   \n",
       "2                     False                       False   \n",
       "3                     False                       False   \n",
       "4                     False                       False   \n",
       "\n",
       "   occupation_Prof-specialty  occupation_Protective-serv  occupation_Sales  \\\n",
       "0                      False                       False             False   \n",
       "1                      False                       False             False   \n",
       "2                      False                       False             False   \n",
       "3                      False                       False             False   \n",
       "4                       True                       False             False   \n",
       "\n",
       "   occupation_Tech-support  occupation_Transport-moving  income_<=50K  \\\n",
       "0                    False                        False          True   \n",
       "1                    False                        False          True   \n",
       "2                    False                        False          True   \n",
       "3                    False                        False          True   \n",
       "4                    False                        False          True   \n",
       "\n",
       "   income_>50K  \n",
       "0        False  \n",
       "1        False  \n",
       "2        False  \n",
       "3        False  \n",
       "4        False  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dummies.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this example, we called *get_dummies* on a DataFrame containing both the training and the test data. This is important to ensure categorical values are represented in the same way in the training set and the test set. \n",
    "We could have done it separetely, but in that case we would need to ensure that both datasets contain the same categories for all features."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Sklearn* has also a function for one-hot encoding. *OneHotEncoder* assumes all features are categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us select only categorical features using `select_dtypes` method which returns a subset of the DataFrame’s columns based on the column dtypes. We can define with *include* or *exclude* a list of data types to be included or excluded."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['workclass', 'education', 'capital-gain-category', 'occupation'], dtype='object')"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X=data.drop(columns='income')\n",
    "y=data['income']\n",
    "categorical=X.select_dtypes(exclude='number').columns\n",
    "categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['age', 'hours-per-week'], dtype='object')"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "numerical= X.select_dtypes(include='number').columns\n",
    "numerical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can now quickly visualize the different values that each categorical variable can take:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for  categorical_variable in categorical:\n",
    "    X[categorical_variable].value_counts().plot(kind='bar', figsize=(4,3), title=categorical_variable)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*OneHotEncoder* has a parameter *drop* with which we can specify whether we want to drop one of the dummy variables in order to avoid collinearity.  We could do this with *OneHotEncoder(drop ='first')*.  Note that this is useful in situations where perfectly collinear features cause problems, such as when feeding the resulting data into an unregularized linear regression model. (When features are highly correlated, small changes in the data can cause large fluctuations in the regression coefficients, and also it becomes harder to determine the importance of each feature). However, dropping one category breaks the symmetry of the original representation and can therefore induce a bias in downstream models, for instance for penalized linear classification or regression model or decision trees.\n",
    "Let's fit *OneHotEncoder* to categorical variables only.\n",
    "We will also change the default value of the parameter *sparse_output* which defines whether the encoder will return a sparse matrix or an array. Since we will later convert the output of the encoder to pandas dataframe, we want the output to be an array."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array(['?', 'Federal-gov', 'Local-gov', 'Never-worked', 'Private',\n",
       "        'Self-emp-inc', 'Self-emp-not-inc', 'State-gov', 'Without-pay'],\n",
       "       dtype=object),\n",
       " array(['10th', '11th', '12th', '1st-4th', '5th-6th', '7th-8th', '9th',\n",
       "        'Assoc-acdm', 'Assoc-voc', 'Bachelors', 'Doctorate', 'HS-grad',\n",
       "        'Masters', 'Preschool', 'Prof-school', 'Some-college'],\n",
       "       dtype=object),\n",
       " array(['cat1', 'cat2', 'cat3', 'cat4'], dtype=object),\n",
       " array(['?', 'Adm-clerical', 'Armed-Forces', 'Craft-repair',\n",
       "        'Exec-managerial', 'Farming-fishing', 'Handlers-cleaners',\n",
       "        'Machine-op-inspct', 'Other-service', 'Priv-house-serv',\n",
       "        'Prof-specialty', 'Protective-serv', 'Sales', 'Tech-support',\n",
       "        'Transport-moving'], dtype=object)]"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ohe = OneHotEncoder(drop='first', sparse_output=False)\n",
    "ohe.fit(X[categorical])\n",
    "ohe.categories_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The attribute *categories_* lists the categories of each feature determined during fitting (in order of the features in X and corresponding with the output of transform). This includes the category specified in drop (if any)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's do `transform`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['workclass_Federal-gov', 'workclass_Local-gov',\n",
       "       'workclass_Never-worked', 'workclass_Private',\n",
       "       'workclass_Self-emp-inc', 'workclass_Self-emp-not-inc',\n",
       "       'workclass_State-gov', 'workclass_Without-pay', 'education_11th',\n",
       "       'education_12th', 'education_1st-4th', 'education_5th-6th',\n",
       "       'education_7th-8th', 'education_9th', 'education_Assoc-acdm',\n",
       "       'education_Assoc-voc', 'education_Bachelors',\n",
       "       'education_Doctorate', 'education_HS-grad', 'education_Masters',\n",
       "       'education_Preschool', 'education_Prof-school',\n",
       "       'education_Some-college', 'capital-gain-category_cat2',\n",
       "       'capital-gain-category_cat3', 'capital-gain-category_cat4',\n",
       "       'occupation_Adm-clerical', 'occupation_Armed-Forces',\n",
       "       'occupation_Craft-repair', 'occupation_Exec-managerial',\n",
       "       'occupation_Farming-fishing', 'occupation_Handlers-cleaners',\n",
       "       'occupation_Machine-op-inspct', 'occupation_Other-service',\n",
       "       'occupation_Priv-house-serv', 'occupation_Prof-specialty',\n",
       "       'occupation_Protective-serv', 'occupation_Sales',\n",
       "       'occupation_Tech-support', 'occupation_Transport-moving'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_cat= ohe.transform(X[categorical])\n",
    "encoder_feature_names = ohe.get_feature_names_out(categorical)\n",
    "encoder_feature_names"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's now convert the array to pandas dataframe:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_cat= pd.DataFrame(X_cat, columns = encoder_feature_names)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's now concatenate the transformed categorical with unchanged numerical variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32561, 42)\n"
     ]
    }
   ],
   "source": [
    "X_num = X[numerical]\n",
    "X_enc = pd.concat([X_num, X_cat], axis = 1)\n",
    "print(X_enc.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us now see how the encoded data frame looks like. As we saw we now have 40 columns. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>workclass_Federal-gov</th>\n",
       "      <th>workclass_Local-gov</th>\n",
       "      <th>workclass_Never-worked</th>\n",
       "      <th>workclass_Private</th>\n",
       "      <th>workclass_Self-emp-inc</th>\n",
       "      <th>workclass_Self-emp-not-inc</th>\n",
       "      <th>workclass_State-gov</th>\n",
       "      <th>workclass_Without-pay</th>\n",
       "      <th>education_11th</th>\n",
       "      <th>education_12th</th>\n",
       "      <th>education_1st-4th</th>\n",
       "      <th>education_5th-6th</th>\n",
       "      <th>education_7th-8th</th>\n",
       "      <th>education_9th</th>\n",
       "      <th>education_Assoc-acdm</th>\n",
       "      <th>education_Assoc-voc</th>\n",
       "      <th>education_Bachelors</th>\n",
       "      <th>education_Doctorate</th>\n",
       "      <th>education_HS-grad</th>\n",
       "      <th>education_Masters</th>\n",
       "      <th>education_Preschool</th>\n",
       "      <th>education_Prof-school</th>\n",
       "      <th>education_Some-college</th>\n",
       "      <th>capital-gain-category_cat2</th>\n",
       "      <th>capital-gain-category_cat3</th>\n",
       "      <th>capital-gain-category_cat4</th>\n",
       "      <th>occupation_Adm-clerical</th>\n",
       "      <th>occupation_Armed-Forces</th>\n",
       "      <th>occupation_Craft-repair</th>\n",
       "      <th>occupation_Exec-managerial</th>\n",
       "      <th>occupation_Farming-fishing</th>\n",
       "      <th>occupation_Handlers-cleaners</th>\n",
       "      <th>occupation_Machine-op-inspct</th>\n",
       "      <th>occupation_Other-service</th>\n",
       "      <th>occupation_Priv-house-serv</th>\n",
       "      <th>occupation_Prof-specialty</th>\n",
       "      <th>occupation_Protective-serv</th>\n",
       "      <th>occupation_Sales</th>\n",
       "      <th>occupation_Tech-support</th>\n",
       "      <th>occupation_Transport-moving</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>40</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50</td>\n",
       "      <td>13</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>40</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53</td>\n",
       "      <td>40</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28</td>\n",
       "      <td>40</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  hours-per-week  workclass_Federal-gov  workclass_Local-gov  \\\n",
       "0   39              40                    0.0                  0.0   \n",
       "1   50              13                    0.0                  0.0   \n",
       "2   38              40                    0.0                  0.0   \n",
       "3   53              40                    0.0                  0.0   \n",
       "4   28              40                    0.0                  0.0   \n",
       "\n",
       "   workclass_Never-worked  workclass_Private  workclass_Self-emp-inc  \\\n",
       "0                     0.0                0.0                     0.0   \n",
       "1                     0.0                0.0                     0.0   \n",
       "2                     0.0                1.0                     0.0   \n",
       "3                     0.0                1.0                     0.0   \n",
       "4                     0.0                1.0                     0.0   \n",
       "\n",
       "   workclass_Self-emp-not-inc  workclass_State-gov  workclass_Without-pay  \\\n",
       "0                         0.0                  1.0                    0.0   \n",
       "1                         1.0                  0.0                    0.0   \n",
       "2                         0.0                  0.0                    0.0   \n",
       "3                         0.0                  0.0                    0.0   \n",
       "4                         0.0                  0.0                    0.0   \n",
       "\n",
       "   education_11th  education_12th  education_1st-4th  education_5th-6th  \\\n",
       "0             0.0             0.0                0.0                0.0   \n",
       "1             0.0             0.0                0.0                0.0   \n",
       "2             0.0             0.0                0.0                0.0   \n",
       "3             1.0             0.0                0.0                0.0   \n",
       "4             0.0             0.0                0.0                0.0   \n",
       "\n",
       "   education_7th-8th  education_9th  education_Assoc-acdm  \\\n",
       "0                0.0            0.0                   0.0   \n",
       "1                0.0            0.0                   0.0   \n",
       "2                0.0            0.0                   0.0   \n",
       "3                0.0            0.0                   0.0   \n",
       "4                0.0            0.0                   0.0   \n",
       "\n",
       "   education_Assoc-voc  education_Bachelors  education_Doctorate  \\\n",
       "0                  0.0                  1.0                  0.0   \n",
       "1                  0.0                  1.0                  0.0   \n",
       "2                  0.0                  0.0                  0.0   \n",
       "3                  0.0                  0.0                  0.0   \n",
       "4                  0.0                  1.0                  0.0   \n",
       "\n",
       "   education_HS-grad  education_Masters  education_Preschool  \\\n",
       "0                0.0                0.0                  0.0   \n",
       "1                0.0                0.0                  0.0   \n",
       "2                1.0                0.0                  0.0   \n",
       "3                0.0                0.0                  0.0   \n",
       "4                0.0                0.0                  0.0   \n",
       "\n",
       "   education_Prof-school  education_Some-college  capital-gain-category_cat2  \\\n",
       "0                    0.0                     0.0                         1.0   \n",
       "1                    0.0                     0.0                         0.0   \n",
       "2                    0.0                     0.0                         0.0   \n",
       "3                    0.0                     0.0                         0.0   \n",
       "4                    0.0                     0.0                         0.0   \n",
       "\n",
       "   capital-gain-category_cat3  capital-gain-category_cat4  \\\n",
       "0                         0.0                         0.0   \n",
       "1                         0.0                         0.0   \n",
       "2                         0.0                         0.0   \n",
       "3                         0.0                         0.0   \n",
       "4                         0.0                         0.0   \n",
       "\n",
       "   occupation_Adm-clerical  occupation_Armed-Forces  occupation_Craft-repair  \\\n",
       "0                      1.0                      0.0                      0.0   \n",
       "1                      0.0                      0.0                      0.0   \n",
       "2                      0.0                      0.0                      0.0   \n",
       "3                      0.0                      0.0                      0.0   \n",
       "4                      0.0                      0.0                      0.0   \n",
       "\n",
       "   occupation_Exec-managerial  occupation_Farming-fishing  \\\n",
       "0                         0.0                         0.0   \n",
       "1                         1.0                         0.0   \n",
       "2                         0.0                         0.0   \n",
       "3                         0.0                         0.0   \n",
       "4                         0.0                         0.0   \n",
       "\n",
       "   occupation_Handlers-cleaners  occupation_Machine-op-inspct  \\\n",
       "0                           0.0                           0.0   \n",
       "1                           0.0                           0.0   \n",
       "2                           1.0                           0.0   \n",
       "3                           1.0                           0.0   \n",
       "4                           0.0                           0.0   \n",
       "\n",
       "   occupation_Other-service  occupation_Priv-house-serv  \\\n",
       "0                       0.0                         0.0   \n",
       "1                       0.0                         0.0   \n",
       "2                       0.0                         0.0   \n",
       "3                       0.0                         0.0   \n",
       "4                       0.0                         0.0   \n",
       "\n",
       "   occupation_Prof-specialty  occupation_Protective-serv  occupation_Sales  \\\n",
       "0                        0.0                         0.0               0.0   \n",
       "1                        0.0                         0.0               0.0   \n",
       "2                        0.0                         0.0               0.0   \n",
       "3                        0.0                         0.0               0.0   \n",
       "4                        1.0                         0.0               0.0   \n",
       "\n",
       "   occupation_Tech-support  occupation_Transport-moving  \n",
       "0                      0.0                          0.0  \n",
       "1                      0.0                          0.0  \n",
       "2                      0.0                          0.0  \n",
       "3                      0.0                          0.0  \n",
       "4                      0.0                          0.0  "
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_enc.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "OneHotEncoder supports aggregating infrequent categories into a single output for each feature. The parameters to enable the gathering of infrequent categories are *min_frequency* and *max_categories*.\n",
    "\n",
    "- *min_frequency* is either an integer greater or equal to 1, or a float in the interval (0, 1). If *min_frequency* is an integer, categories with a cardinality smaller than *min_frequency* will be considered infrequent. If *min_frequency* is a float, categories with a cardinality smaller than this fraction of the total number of samples will be considered infrequent. The default value is 1, which means every category is encoded separately.\n",
    "\n",
    "- *max_categories* is either None or any integer greater than 1. This parameter sets an upper limit to the number of output features for each input feature. *max_categories* includes the feature that combines infrequent categories."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's set for our one-hot encoder that if a category appears in less than 1% of all the data, it should be considered infrequent."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['workclass_?', 'workclass_Federal-gov', 'workclass_Local-gov',\n",
       "       'workclass_Private', 'workclass_Self-emp-inc',\n",
       "       'workclass_Self-emp-not-inc', 'workclass_State-gov',\n",
       "       'workclass_infrequent_sklearn', 'education_10th', 'education_11th',\n",
       "       'education_12th', 'education_5th-6th', 'education_7th-8th',\n",
       "       'education_9th', 'education_Assoc-acdm', 'education_Assoc-voc',\n",
       "       'education_Bachelors', 'education_Doctorate', 'education_HS-grad',\n",
       "       'education_Masters', 'education_Prof-school',\n",
       "       'education_Some-college', 'education_infrequent_sklearn',\n",
       "       'capital-gain-category_cat1', 'capital-gain-category_cat2',\n",
       "       'capital-gain-category_cat4',\n",
       "       'capital-gain-category_infrequent_sklearn', 'occupation_?',\n",
       "       'occupation_Adm-clerical', 'occupation_Craft-repair',\n",
       "       'occupation_Exec-managerial', 'occupation_Farming-fishing',\n",
       "       'occupation_Handlers-cleaners', 'occupation_Machine-op-inspct',\n",
       "       'occupation_Other-service', 'occupation_Prof-specialty',\n",
       "       'occupation_Protective-serv', 'occupation_Sales',\n",
       "       'occupation_Tech-support', 'occupation_Transport-moving',\n",
       "       'occupation_infrequent_sklearn'], dtype=object)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ohe_inf = OneHotEncoder(min_frequency=0.01, sparse_output=False)\n",
    "ohe_inf.fit(X[categorical])\n",
    "ohe_inf.get_feature_names_out()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see from the previous output that for the feature occupation we have now a category called 'occupation_infrequent_sklearn'. We can check which categories are aggregated in this category:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array(['Never-worked', 'Without-pay'], dtype=object),\n",
       " array(['1st-4th', 'Preschool'], dtype=object),\n",
       " array(['cat3'], dtype=object),\n",
       " array(['Armed-Forces', 'Priv-house-serv'], dtype=object)]"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ohe_inf.infrequent_categories_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that two categories 'Armed-Forces' and 'Priv-house-serv' are grouped. Checking the presence of these categories in the data we can confirm that it is less than 1%."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "occupation\n",
       "Prof-specialty       0.127146\n",
       "Craft-repair         0.125887\n",
       "Exec-managerial      0.124873\n",
       "Adm-clerical         0.115783\n",
       "Sales                0.112097\n",
       "Other-service        0.101195\n",
       "Machine-op-inspct    0.061485\n",
       "?                    0.056601\n",
       "Transport-moving     0.049046\n",
       "Handlers-cleaners    0.042075\n",
       "Farming-fishing      0.030527\n",
       "Tech-support         0.028500\n",
       "Protective-serv      0.019932\n",
       "Priv-house-serv      0.004576\n",
       "Armed-Forces         0.000276\n",
       "Name: proportion, dtype: float64"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X['occupation'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These levels were grouped together not because they are similar, but because we don't really have enough data to say much about them."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we fitted the *OneHotEncoder* on the training data, as we should do, if some category is very rare, it might appear only when we transform test data. The behavior for this case is handled with the parameter *handle_unknown*:\n",
    "- error: This is the default behavior, and an error is raised if an unknown category is present during transform.\n",
    "- ignore: When an unknown category is encountered during transform, the resulting one-hot encoded columns for this feature will be all zeros. This means that unknown categories will have the same mapping as the dropped category. For this case, we should not drop any category with the *drop* parameter (as the dropped category is represented with all the zeros of the dummy variables). \n",
    "- infrequent_if_exist: When an unknown category is encountered during transform, the resulting one-hot encoded columns for this feature will map to the infrequent category if it exists. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Ordinal encoding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Categorical ordinal variables have categories that follow a logical ordering. Some examples of ordinal data include:\n",
    "- Socioeconomic status (low income, middle income or high income)\n",
    "- Education level (high school, bachelor’s degree, master’s degree or PhD)\n",
    "- Satisfaction rating (extremely dislike, dislike, neutral, like or extremely like).\n",
    "\n",
    "Ordinal variables are encoded using scikit-learn *OrdinalEncoder*. In order to use *OrdinalEncoder*, we have to first specify the order in which we would like to encode our ordinal variable. In our case let's encode education level with ordinal encoder (let's assume some ranking between the education levels), and capital-gain-category."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import OrdinalEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "education\n",
       "HS-grad         10501\n",
       "Some-college     7291\n",
       "Bachelors        5355\n",
       "Masters          1723\n",
       "Assoc-voc        1382\n",
       "11th             1175\n",
       "Assoc-acdm       1067\n",
       "10th              933\n",
       "7th-8th           646\n",
       "Prof-school       576\n",
       "9th               514\n",
       "12th              433\n",
       "Doctorate         413\n",
       "5th-6th           333\n",
       "1st-4th           168\n",
       "Preschool          51\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X['education'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "capital-gain-category\n",
       "cat1    29849\n",
       "cat2     1942\n",
       "cat4      613\n",
       "cat3      157\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X['capital-gain-category'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array(['Preschool', '1st-4th', '5th-6th', '7th-8th', '9th', '10th',\n",
       "        '11th', '12th', 'HS-grad', 'Prof-school', 'Assoc-acdm',\n",
       "        'Assoc-voc', 'Some-college', 'Bachelors', 'Masters', 'Doctorate'],\n",
       "       dtype=object),\n",
       " array(['cat1', 'cat2', 'cat3', 'cat4'], dtype=object)]"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categories=[['Preschool','1st-4th','5th-6th','7th-8th','9th','10th','11th','12th','HS-grad','Prof-school','Assoc-acdm','Assoc-voc','Some-college','Bachelors','Masters','Doctorate'],\n",
    "            ['cat1','cat2','cat3','cat4']]\n",
    "encoder = OrdinalEncoder(categories=categories)\n",
    "encoder.fit(X[['education','capital-gain-category']] )\n",
    "encoder.categories_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32561, 6)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_enc = X.copy()\n",
    "X_enc[['education','capital-gain-category']] = encoder.transform(X[['education','capital-gain-category']]  )\n",
    "X_enc.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>capital-gain-category</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>occupation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>13.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>40</td>\n",
       "      <td>Adm-clerical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>13</td>\n",
       "      <td>Exec-managerial</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>40</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53</td>\n",
       "      <td>Private</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>40</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>40</td>\n",
       "      <td>Prof-specialty</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age         workclass  education  capital-gain-category  hours-per-week  \\\n",
       "0   39         State-gov       13.0                    1.0              40   \n",
       "1   50  Self-emp-not-inc       13.0                    0.0              13   \n",
       "2   38           Private        8.0                    0.0              40   \n",
       "3   53           Private        6.0                    0.0              40   \n",
       "4   28           Private       13.0                    0.0              40   \n",
       "\n",
       "          occupation  \n",
       "0       Adm-clerical  \n",
       "1    Exec-managerial  \n",
       "2  Handlers-cleaners  \n",
       "3  Handlers-cleaners  \n",
       "4     Prof-specialty  "
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_enc.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that here we remained with the same number of features as we had initially."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((32561, 6), (32561, 6))"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape, X_enc.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Target Encoding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When we have categorical features with high cardinality, using one-hot encoding would inflate the feature space making it more computationally expensive for modeling. A classical example of high cardinality categories are location based such as zip code or region. In this case we can use Target Encoding, where each category is encoded based on the average target values for observations belonging to the category, more specifically the encoding scheme mixes the global target mean with the target mean conditioned on the value of the category."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check again how many categories does each categorical feature have using `nunique()` pandas method:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "workclass has 9 categories\n",
      "education has 16 categories\n",
      "capital-gain-category has 4 categories\n",
      "occupation has 15 categories\n"
     ]
    }
   ],
   "source": [
    "for categorical_variable in categorical:\n",
    "    print(f'{categorical_variable} has {X[categorical_variable].nunique()} categories')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's see how would we encode workclass and occupation with target encoding."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "from  sklearn.preprocessing import TargetEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32561, 6)"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "enc = TargetEncoder(target_type='binary')\n",
    "X_enc = X.copy()\n",
    "X_enc[['workclass','occupation']] = enc.fit_transform(X[['workclass','occupation']], y)\n",
    "X_enc.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can check what are the values that the categories were mapped to with the following: For feature i, encodings_[i] are the encodings matching the categories listed in categories_[i]."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array(['?', 'Federal-gov', 'Local-gov', 'Never-worked', 'Private',\n",
       "        'Self-emp-inc', 'Self-emp-not-inc', 'State-gov', 'Without-pay'],\n",
       "       dtype=object),\n",
       " array(['?', 'Adm-clerical', 'Armed-Forces', 'Craft-repair',\n",
       "        'Exec-managerial', 'Farming-fishing', 'Handlers-cleaners',\n",
       "        'Machine-op-inspct', 'Other-service', 'Priv-house-serv',\n",
       "        'Prof-specialty', 'Protective-serv', 'Sales', 'Tech-support',\n",
       "        'Transport-moving'], dtype=object)]"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "enc.categories_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([0.10406847, 0.38626183, 0.29476285, 0.        , 0.21867381,\n",
       "        0.55696537, 0.28490785, 0.27193089, 0.        ]),\n",
       " array([0.10367319, 0.13450071, 0.11845551, 0.22664396, 0.48393209,\n",
       "        0.11576456, 0.06281553, 0.12490973, 0.04159133, 0.00676869,\n",
       "        0.44896578, 0.32495995, 0.26930666, 0.30487686, 0.20039788])]"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "enc.encodings_"
   ]
  },
  {
   "attachments": {
    "546e8541-dffc-45db-b3e6-05bfbbbd6f4e.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvIAAAGACAYAAADPmYYbAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAAFiUAABYlAUlSJPAAAABhaVRYdFNuaXBNZXRhZGF0YQAAAAAAeyJjbGlwUG9pbnRzIjpbeyJ4IjowLCJ5IjowfSx7IngiOjc1NCwieSI6MH0seyJ4Ijo3NTQsInkiOjM4NH0seyJ4IjowLCJ5IjozODR9XX1TVghhAABijUlEQVR4Xu3dC1wU5f4G8IdTJ7WjpWUFZhZeKihLvKSLqblm6SKWu2kCXlmPKaCWoF3ALgpdEKxQtBLUNNBUVstAj8b6Lw1MTEyNtcSwtCCzpLQjWh3+7+y+6i53hMUdfL6fz7Q7v5lZiNedeeadd2fdSgUQEREREZGq/EM+EhERERGRijDIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqZBbqSCfE5GL2L59u3xGDen++++Xz2ouICAAq1atknPUEJKSkmA0GuUcEdXFoR9+k8+oIXW4+Vr5rG7qJcg/F/uufEYNJXT0YNzscaOcq5ldu3Zhy5Ytco4ayrPPPiuf1VxCQgKmTZsm56ghKMFQCYi1pQT5nTt34vHHH5cVcqZXXnnlooO8si01rEcffRReXl5yjlzR+5nfYPveH+UcNYRhfTtA2+0WOVc39RLkQ+Mz8Y9/XCHnqCE8M9L7ooJ8jx495Bw1hDvvvBMWi0XO1dy5IP/RRx/JCjmTEgqHDBlSpyB/6NAhWSFnyc/PR6dOnS46yLdr1w5HjhyRc9QQ8vLyGORdnBLkd339Mx7S3C4r5Ewf/t9Xrhnku911K3reXT+/FFVuX34hPss9XKcgv3btWgwcOFBWyVnGjRtnDfF1CfKFhYVwd3eXVXIW5W/MIO/66iPI9+vXD4mJibJCzrJ+/XqMHTuWQV4FzgX5iYZeskLOUnL2T7yzdofrBvk+Pp6yQs7yheVIvQR5g8Egq+Qser2eQV4lGOTVob6C/IoVK2SFnGX58uUM8irBIN9wnBHkedcaIiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnuoxlz3GDm1v1k2FlkdyiDnbEWF8rZoecp1rIRkyZNql4MiD1qNykDopWGurttYjoMle4EmHuV+L2KqawdbU8xnzxsthuODYUyvlK7cCiKl+/ABumtK39z3chqgjyP68bXq7RHaeaNKYj22u+jFw5X5lq1zuyEuH31P7nN2a2EFBRyCgzzcmWW9SFDDj18lqXH/cucYiLvTBFjvYRVR8EPutY13k2tW1Al4g7fOzaIy42EoFdRLlLICId6jp4NrdtQZfI0VQYyu7rKpxixN6r7njSRWrhPfw5PPN8bIVTv3aX4BjzVwE2R45G+Br1hniFinrku8B/asX/AJ55fjDaXi1Xa0hH1uGlSaOx4ZicJ6umnjq7YCGmyTprXTfZrqZMXdytdbp0PP3DET7jwhQ22FOpwi/EsW7s1dK2QV30ikRpaSkie8l5qgVP6OzaI3xGGPw6iXInP4Q51I3Q1ENTuQekibZKQ2BbWaCaay7ayn4/FxsK6x5wcKhdTZl8xOkZ0eWjrTYEwSHTK5yGd6uHHVct/L5/JV59vA/CktV/iVhFQd4TD4yv+B9AcEgwfK6VqzWEv4qRt24mxvgNR8oXskbntexltAsWYhrja637jrGrKZO/EhrrSoNIEQ5LZ2nkPDmLrecvBhk74kWIdINbJ19EbSmxLTyejcQQA3yVutLb6OGNAcHxyD5uW2xVbmiN7WqKYaUFlpUR8OvsYV3eXjMB8duL5TpUe0VIHSHaYE4Gsuf6ob31bxoF8ynb0pIDJsQED4C3h2wr0Y6GmamwyOWKcr281rYT8wcsSJ3pJ7dtD9+ybUxiB6iB0X4/NyMI1j2gJsiupkw6cVSrO550UaPzVxF2JE/Fvx9oax110fsBf7wkAvfvcnHlSnBIZDPbdm0xePJC5P0hF9krXInnHhyNJT9q8MLHG/CULKtVoxwj//uBdVj05P3QWofddMLjT76MzfkycFTl1x1Ycm47jT9e3VyAM3KRvZ83/BuPTp6Hoz0TsD51tqzSRTkXELabMEEJciIAGpZa5MJiZM8Pg0HT3hY43DzgrS0b8soOrTkXYrJRvD0eE+S2Hp39ECECYw3+FVCVFsA42gzf2Eykzw6Dn09T4JQZEZ19EX2gFYJEPdOcjpTpvji9NAK+/vHVDl8rSDDCb3ExdLNTkGlaAG2zdET0CUJyvlyBLs5CI4K2+SJOtEd0iB+6KkNu8pNh8DIg+ZQWUUuUtkpDkn8rZM0NgnZmhnjHVaUAC8b7IblYh+jUTKQlaNFso2jjMcliCV0ceTK7Ihumid7wUPZxI5Jh2wOWwLIuBhO0Sl3Z/yknZIZy+7GyJ13nTrizxcl1fLCv9URO2a/6lTlZI3JNBdjwVHeMidyNG0fPx/K0DXhmaFN8Enk/Bk9ZiapGkOW9/aAI76k40282FqctRkinrQh/JhFfyeUXNEXbkBXYuGkNgu5uJWvq1eiC/JkvXhZnY8Px3q9aPJW6BcuXPIWORxYi7P4HseiLKmLcrxvx0gP349XPPTBqidhu1mCcfHsiXjdXcGhr4ongRfuwfnEIvBvySkCjlYXw4XFoOj0ZmQvDYdR4iVoJzDO94PuyBa0C40TgEMExVYSG08ki5OkQv8e2ZaUywqEbnwXPqQuQmZGCMC8L4gO1iN7KKF83RTAkpCFymBa6gEBoWovTra0ZyGquR/zyJISKura/DoEzkpCyRCdO1LJgqWbsbm6JFikfym2HhSJpeTz0yEBGjrrHLV5yRQYsSI2EXmmP0RooF65zt6SioFc0UpZHInCw0lZ6GOelIeVZdxQtypUBsjK5KBmYgvR3QsVraqGfItp4nh7YmIFsjs+uk6yZesQ1DUeyORHho32h7AELFhvgrU/Gif5Roi5OukxJ0N+QZd2PRWyo7opVOsL9g5DlGYYFyol1iBcsysnay2Z2ZpBL+z3jFYSvccfkDz5GjHEYevUeDP/pa7A+dTqwZgbWfFbJv+Bf1yHlhR3wmbkWi18KRj+53eK+JdgsVznPYxieeT4AHRpJfmtkQX4PUp5+Hrn+K/D+8tnw1/ZHL10IYt7fghe0O/B6fGqlZ3N5q2YhBQGIX70GwTq53fI5aGvZKte44AZdLJ4Z5oVr5DzVVRG8n0/GgvE6EeSM0N0pSsfNyNjWDPp5KUiaoheBQwmO4UhakQQdspFlqSbk7fBA2BYROAPEaw4OROTqFES7FyFme3X9w1Q1PXw7O34oqaV/HLIOlr+07+mlDCgoxunqkoO/HzT2H9Bs64mu4sGUz37eOhnua+uFt+MzORN52ZHQODRhU3h3VtrqRLVtpXtII9a+wN3T2lIoYJCvk6LOIqwniH2fcmLlr8T4XGSsKYBmdgpSZgWKunKSa0Sc2I9Fiv1Y4p6qT7mUnn6PkEykWbcVJ3KzxMnabHGy9kp2tVfIiJxl82TbUJmy04U7xhQj99MlQL8QDO/peJy5RhsI491FWJSZXeFIiTP7srEG/fHYsF5oImuKtkNDECyfN1YqCvLrEO5T/h+Awx1r9pvx4X4geNQwOGSKK73gP0o0pXklsivMBgX46pM9wGPBeOgWWVL8qxeGj+svZ8iZtD7KwctOax3isr9FWkCZj4N19LaONy0+ddo2X5lhBmhuk8+tPOHZRzzsKxCnDeQUp4pRsMcM87pkxM+cAN/RUaJoRlF1Y6j/KR+pwZQUF8GyQ5wsr0xEVIgftNNNoireG9W0VbMr5ROqX326WnvhL/BB6OY8ZM1yPHFCU294K/uxk6er6VnXw9DHcQS+Z0dlz7mbJ110yVR215oLd6wpwe+/iIe7PHGDrWDHHTco/6SPFFU4Vv73X5Vw54Nb2tjmz2vSCjfeLZ83UioK8pXdtcbujjVnSpAnzsg6VHAbo2tuUlLdVhz/1Tbv6Ccc/wTo1f42hzM5xTU3lgmY1LD+KkHx4VyYt5qQPDcCEzRBsMbDomriuAgczeRTcrJTuUgc0R5uLVqhvU8QwuanIuuUO4xjjHIFchmHMxCh9UCzVh7w1kxA9FITiprrYBytlSuQaxH7vyILsrdmIHV+FMKGahG+RpQPF1XzeQaBJ13kYiq7a41z71hzLVpcL582UioK8pXdtca5d6xp0qT8SQE1hBLkzjeg/T+boZVnVwRNTURqTgncJxrBeOhachcZEbbGBwtyT6C0tBB55kykLYxGUH3cspLqUTFMs/0QfyII6UdOi7b6FlmbM5EUGwq9citLcikFGyIwwEPs/zy84TsxGskbitB0oBFBA+UKRI1OU1yjhO6vCvCzrWCnCD8rne6dKuqtB665Tumuz8WRH23z5505jEOfyOeNVOMaIy9Ctze24tD35S86/v7TYfHfYWh7o23e0U1o3Q/Y8e3hcmOvfj7CEYWXxJ5EGKea4LNwN078WYrCfZnIXL0A0QG2D+2RqyiCJUe8R4YZoOti3zLFMG9Jkc/JNVhgWSoe/EVbtbXroPjLgox1ZjlDLuG4CTFD43FidDoKT5ei9GAWMjcnIW6KvswQHKLGpCV8+gYDnyzEms8dc9zv5lQk73dHUHflQ3TlNenSH0E3bkXKmh0OOe5oxhJskM8bq8YV5O/WYujdwJL31jl+qFUcqDa8twTo1h/e9mPgz/NE14d6AQsXYvMRWVL8sQMbVpb/sCs5X5ElS5xb62EY7IOWdpeIlTukMB66End4aTTAunBMmJ6MjK3KGPlEhOm6wriuGb/wxqV4wWeyaJE5QTDMNcEs2ipjZQyCfLSIP1gfdzSnepNvQbJ40A3Vwd3+ovCBDJi2yOdEjdA1umcRP7wIix55EJHJ67Djs43YMG84Hg2cBwyfC6O2kq68awfD+FIAfp73GP4dIbLcZ1uxeeFwjJm8soLe/calcQV5dEHQa7Phs2E0Hh/zPDaYt2JHxkJEPj4QL5l74anZwegg1yyrw8g4PNVtHcJHDMeidRuxw7wEr455DK/zC58uCXcvX2hgQnhwBJI3inC41YRE5V7YwSY0Yzp0KT5PZSAr1g+nN0yAn3YAJsSa0XR4CnK3LYBBLM+y8O4zrqEldKJtUmZ4wTLPgAHaIEQsLkDX2CzsXhMKH2Qh72B1txiiBtHRB6FiPxcz3oD4dcr+LwOp4gTMu388LB3lOkSNkif8X9+F5TFdcWzFFIwx+OPVD0vw0Bu7sHF+gOONTMpoO2wF1qdNxy2W1xFmGIhX01th1AdrMFkub6waWZAHmnR7Dov/bw1GXWfG64EDMSb4deTfMh2Lcz7G5G5VjHf/Vy9MTt2H+MEl+PAFf4x58nUcuncuFr86TK5ADapLODK2xcHvtAkTdAMwYGIczM31SMnNwgKlSbLz+CU0TlDVt0RWvqwlNDOSkHWwVCwvxbfZaYgbr4G7uw4LxHz6eNnb2yvSujyyl2228m/l5bf1Vs8dgavF32h1YAVXPapY1twLgbHpyCtU2kr5PEMSwgd7oqV4v+0W89H9bfvIcm1dru2kyupUN8pdu7amINzLgji92P8FRiC5oCvitu1GWogPsC0PeTznIrXwCMCCor/EsbuGvXBXuqOXMUFkuaP4Rmz32f9twDMjuzje8ltkvW+K1sDfQ85LN/Sejpj0g9btzOnvILjnMDxV5c8W2a82v5sLchM74VL5/KKFxmei2123oo8PL8862xeWI/gs97D4R+2Nmz0qHPBfqV27dqFHjx5Yu3YtDAalr5ScSa/Xw2KxWKfaSkhIwLRp01BYWCgCsXp3MGqh/I2HDBmCpKQkWam5gIAA7Ny5E4cOHZIVcpb8/Hx06tTJ2k5GY+0/9t6uXTv069cPK1askBVyluXLl2Ps2LHIy8uDlxdH9ruy9zO/wa6vf8ZEA8/Ina3k7J94Z+0ODOvbAdpuFY71rrVG1yNPRERERHQ5YJAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIht1JBPr9oofGZ6HbXrejj4ykr5CxfWI7gs9zDeGakN272uFFWa2bXrl3o0aMH1q5dC4PBIKvkLHq9HhaLxTrVVkJCAqZNm4bCwkK4u7vLKjmL8jceMmQIkpKSZKXmHnroIeTl5WHOnDmyQs4UHByMyMhIREdHy0rNtWvXDv369cOKFStkhZxl+fLlGDt2rPW94eXlJavkit7P/Aa7vv4ZEw29ZIWcpeTsn3hn7Q4M69sB2m63yGrdMMirDIO8etQlyCvtYzKZMHToUFkhZ9q6dSu8vb2xY8cOWam5W2+9Fd9//72co4agBMRly5bJuZpjkG84DPLq8dZ/UnHk2Cnc4t5SVsiZCn74FZ3bX49RfYfLSt3US5Dv0LkP/nV1EzlHzvbH6bP48P0k3OV1u6zUTGb2JiR/9DratGkjK+RsJ0+exNvPrJVzNXcuyFPDuf322/H111/LuZrr06cPDh48iHHjxskKOdNrr72GkJAQJCYmykrNXX311WjVqpWcI2f78ccf8dFHH8HPz09WyBXN37QIxX/tkXPUEG5r8SBG93OhIH/Flf/Etde748orrpQVcpa///4bvx47gv15X19UkM/6ZR3+ddW1skLO9lPhT3ht7LtyruZefvll6/CB/fv344orrpBVcpa+fftah8i89957slJzAQEB2LlzJw4dOiQr5Cz5+fno1KmTdQiU0WiU1Zq76qqr4ObmJufI2c6ePcsgrwIf5L6HAz/l4KHOo2WFnOmj3MXo03EYHrhzkKzUTb0EeWXH2L17d+Tk5MgKOYsyjjc9Pd16QOvQoYOs1syPxd8hefvL6NVBh5uv6yir5Cw78tNx5s8/MPmBl2Sl5jhGvmHVZYw8g3zDqWuQ59CahsOhNeqRsS8V+3/4HEN8JsoKOcvZv0qwIfcdPOj1GDQdBspq3fCuNUREREREKsQgT0RERESkQgzyREREREQqxCBPRERERKRCDPJERERERCrEIE9EREREpEIM8kREREREKsQgT0RERESkQgzyREREREQqxCBPRERERKRCDPJERERERCrEIE9EREREpEIM8kREREREKsQgT0RERESkQgzyREREREQqxCBPRERERKRCDPJERERERCrEIE9EREREpEIM8kREREREKsQgT0RERESkQgzyREREREQqxCBPRERERKRCDPJERERERCrEIE9EREREpEIM8kRELi8bMW5ucKt2MiD1qNykDopWGurttYiIyHkY5ImIXJ47fGLjEHd+ikRgF1HuEohIh7oOns1tWxARUeOnniB/NBWGCnugyk4xyJabXLwipI4QrzUiVTyj2tgz727c5V7V9A72yHVr5jjS/y22m/elnK9MJev99i3SZz+JII3t5/d7IARz132Lk3Lx5S57Ttn3T8WTYWU9vBN2xFhfK2aHnKda8IRuRjjCz09h8Oskyp38EOZQN0LT0rZFXbgHpKG0NA2BbWWBaq5Bj1W8ekIqUJiOaeWygOM0bd1xuXINffGO2O5JpBfK+Up9ibcreP3jny3D84ah6Gf9+Q8j6Mll+PyYXKgy6gnyzcWBzKHnKRQ6pT441K6mTD5wt25Al05fBDwfgRkVTneitVzL6f4Qb+DAoZi5UfxGMxdiSdpCzHwI+GjyUEwSgf+MXO1y5t7F/r0Th8jRPqLqg8BnHes6z6a2DUgFZEfEnAxkz/VDexEa22uiYD5lW1pywISY4AHw9pCBspMvDDNTYZHLFeXCofUkTMwfsCB1pp/ctj18g+ORXcvjb6PHYxVRhbyGT6wgE9imvu2ayLWc7+i6p2EwLMNRzwDMSkvGkiXj0OFgHIIffBrpR+RKKqKeIN9SA6NDz1MQfJW6Jsiupkw6eFo3qAt3BK4uRenqQO5oL0oXDAkZh3EVTn3RUJ18Rze8jYQvHkTs6jfwxLC+6Nm7L/yeW4iUuL7YE/s2PiqQK17GPP3t3zvhCBusvHs84RfiWDf2qodu3l6RKC0tRWQvOU/OtdCIoG2+iDOnIzrED12VITf5yTB4GZB8SouoJZnINKchyb8VsuYGQTszA8W2LStRgAXj/ZBcrEN0aibSErRotjECvmOSxRI6r0GPVbx6QupxszawgkxgmwzdWsi1nOzM5zC9kI4bpr+JRXEBeLB3T/TUBWD26mSMQzpiV3yuuk6+RjpG3vbBMMOKbJgmesPDzQPeI5JhsS4rRvb8MBg07W29Ucoy7QTEb7c/hJUdWnOuhysbxdvjMUFu69HZDxErLSixrkMX4/hnKxE9Rl7eumcoJj+zEnt+lQurcPLzZXje72HrJbmHA+Pw6ZGyb72T+NZyFK11A3HvLbIkte3VR5xqfIrjNfg5dIGtlzYGGTvi4dfJ1pMbtUX+6z+ejcQQA3yVuvK+8vDGgLK9teWG1sj3qXgPWVZGwK+zh3V5e03Z9yNdlCIDFqRGQt9fh8DRGiinYrlbUlHQKxopyyMROFgLbX89jPPSkPKsO4oW5cp9ZGVyUTIwBenvhIrX1EI/JQkp8/TAxgxkc1jHRarqWFUCy7oYTNAqddv7qr3GUO6YU/bqybn3abZ4T8YH+1qvyCjvR78yV12IXNZfx/F5cgwmP/CA9RivDImNTv6yBkNiz+DbdXFyuwfgP3klLH/IRecc/BZ7/gX07XcvHK4B/KsLevqLQ1nBcdUNvW3UH3bNmqlHXNNwJJsTET7aF15i92ee6QXfly1oJcJfpjkT6akR8D2djIg+OsRXN3g7Ixy68VnwnLoAmRkpCPOyID5Qi+itjPIXw3Z5KwaW62yXtxa9NARXfRKDoAeqvrx1cnMMhjwSh8/bBOBNsd2MAb9j2ZMJ+OR3uYJVC/R96UN8ssSv3BWAk99YsEf8a2jScFfyGpEFMI42wzdWvHdmh8HPpylwyoyIzr6IPtAKQaKeaU5HynRfnF4q3lv+8SL+Va0gwQi/xcXQzU5BpmkBtM3SxfsxCMn5cgW6OMN9bb3wdnwmZyIvOxIah5FSTeEt2g84gdPV7Mp0D2nE2he4e3YV/zWhgEG+Tsofq8T7YrEB3vpknOgfJerifWVKgv6GLOsxJ2JDdSe66Qj3D0KWZxgWKO/HEC9YlKsuL5vZ8UQu7gekP/UYgiPzcMPoKNuQ2KFXYVtkEIZMSUdVuxrL28EivH+EM/2mYlHabEzq9DlmPpOKPLnc6u4ALMnej6k95fx53+Lbz8XD9S0cA74KNOogX9RZ7ABFSNApvU7+Ytd43IyMbc2gn5eCpCl6aPtroQsIR9KKJOiQjSxLNR/o2+GBsC1piAzQQTs4EJGrUxDtXoSY7dVFFSrn14/xzmTl8lYKlrxhu7zVd9hEvLlpYTWXtyxIi10pQsprSF48zrrdg8bZWDTzZnz7iVylKn98idREE6AVP/NuWaNaKIIhQbwHhinvnUBoWgPFWzOQ1VyP+OVJCBV1rdIDPCMJKUt04j2TBUs1IS+3RIuUD+W2w0KRtDweemQgI4cfNXeWkuIiWHaI/eHKRESF+EE7XbwnUICiasa7N7tSPqF6Ve5YJU5/M9YUQCNOblNmBYq68t4wIk4ccyLFMSdxT9XXTpSefo+QTKRZtxXvx1lpSJntjqJXsqs9sSa6lE5mvI2Za27AEx8swWzjg7YhsdPfwNrUccCauTB9VsnAF5EpVr7wJbrMfBOLXtKjr9xuUd8z+FiuUpWj65Zh6f7WeMK/JxpokE+9adRBHn26Wns2zmutQ1z2t0gLKDPyvaO3dQxj8anTtvnKDDNAc5t8buUJzz7iYZ84ANoKZJWAIOsnwctOF+5Yc3LX50hDTwQML3N569q+GDLJC8cTtmFPRe/XAgs+3Q8MeexBh572Jj0NCOgnZyrzh2j7Z6Yh4YgfYl/TN9hY/cZFD9/Ojh98bekfh6yD5cfoenpZ31XV9vLC3w8a+57jtp6w9vPmc+R1vTucgQitB5q18oC3ZgKil5pQ1FwH42itXIEuibLHKvggdHMesmY5XgFBU294K8eck6er6VnXw9DHcQS+Z0fl/bibV0/okvl4sm2oTNnpwh1lTiL3UxPQLwD6no794i20fhh/93G8nbmnwk6+M/v2WDOFfphjpmg7NADiFKBKJz9JwNOT09FWnAQ80Vt9l+obd5CvzF8lKD6cC/NWE5LnRmCCJghRomwuqiaOXwk0k0+pKpXdtebCHWvO/PGz+K8X2raxzdu7weNm8d+jFY9h//U4Phdv1vblPuF+DW6o6pNjv32Jt8cE4/lPemK2aTb8yoybp3pwqhgFe8wwr0tG/MwJ8B1tfVdV28uLf8pHcrJimGb7If5EENKPnEZp6bfI2pyJpNhQ6JVbWZILEseqIguyt2YgdX4UwoZqEb5GlA8XVfPBZIFXT8jFVHbXmgt3rDmDk7+Ih7tuxg22gp0bbMf4IxWPYT/5q3KGWkGmaCKyQRVX34+uex6PPf4OMDUZb00v07GoEpdZkC9B7nwD2v+zGVp5dkXQ1ESk5pTAfaIRRrkG1YfK7lrjzDvWXIUmV8unZR1Jx/ODgpBQMhCL0l+DoaMa36ou7FQuEke0h1uLVmjvE4Sw+anIOuUO4xi+q1yLBZal4sHfAF1bu37evyzIWGeWM+QqCjZEYICHOFZ5eMN3YjSSNxSh6UAjggbKFYhUprK71jj3jjXXoMX18qmDk9iTIN5Pkz9F+5fWYMlz6htSc87lFeT3JMI41QSfhbtx4s9SFO7LRObqBYgOsN3RgRpOk38p59sWHP3RNm/v58IfxH/b42YP27yD61qjJz7Ht9+Xvbh2HEe/kk/tnNm/DJP9nsYnnSKwdnUk+rInvt7lLjIibI0PFuSeQGlpIfLMmUhbGI2g+rhlJdUjL/hMdgfmBMEw1wTzVmWMfAyCfLSIP1gfN0KkenPchJih8TgxOh2Fp0tRejALmZuTEDdFX2YIDlFj0sQWur/6Aco1e0c/42dltGWnthV+F02L65RuwgoyxZmjFXx+7gd8GjkJQS+XYODyFCx6wkuVPfHnXFZBvsiShVxl7OBgH7S0u+yofFgvRT6nhtGie08YRCBfuabMFzP99ik+esuC1sYu6CBLDjy7YGA3YNl7Hzt8ev3MF+kwlX2zHknHrMA45PV7DSnLx8HrX7JO9agIlpxc6+dHdF3sg3sxzFv4rnItLaGLNSNlhhcs8wwYoA1CxOICdI3Nwu41ofBBFvIO8p4mLiHfgmTxoBuqg7v9IPkDGTBtkc+JGp0W8OmrBz5ZCdPnjp11J83p1g+jBnRvLyuOmnS5DwE3ls8URzNM+Eg+tzmDPfNmYrJ4g01NX4Koh5ShvOp2WQV5dy9faGBCeHAEkjearWPkE5UvSgk2oRm/+alhXfcgJi7yw8/zghD85Ep8/Nnn+HTdO5g2KATL4IeZk/pWcpmrPfSzp6LLhqdh/Pc7SDeL7VbFYfL4hPMfpLU5iU8T5yL9WHt0ueFnfLxwGZaVmT7lZynrgTu8NBpgXTgmTE9GxlZljHwiwnRdYVzXjF+o5jRVfWldFcuaeyEwNh15hWK59epJEsIHe6Jll3DsFvPR/W2psdyXDFX2ZV78ki/n6OiDUNF4MeMNiF+nHKsykDonCN7942HpKNchaoRa6J5A7PCf8fYjwXg++WN8/tmnSJ/3JB4LXAYMn4Fx2koGwFzbF+NeUjLFNEyOsGWKjxc+CePkdDh8TGv/SsyN/RLo3QH4fGW5XLBsVU3uV+9aLq+hNeJglbEtDn6nTZigG4ABE+Ngbq5HSm4WFgwTy7Pz+A2FDajtsNeQlhYJr19XYo7BiMkvfISzg2dj7f+9VuWHUZt0m4i3xDoPnvkIsYFGzHrrW3i9tBBR/nIFxW+5+HSZ8vb9VryZ4zB3dvnJwi+Eqhc+T2UgK9YPpzdMgJ92ACbEmtF0eApyty2AQSzPsvBdRVQryh3WtqYg3MuCOL04VgVGILmgK+K27UZaiA+wLQ95vHhCjdLN8Ht9LZbEeOPnFdEINoQg9sOzePCNNfhofvnvhbFnyxTj0NayDNNEppibfg0CPngDT8jlCstn6bZOv89MSKggF8zNPFrhXXFcmVup0p1SR8o3znXv3h05OTmyQs4yZMgQpKenIz8/Hx06VDj4pFI/Fn+H5O0vo1cHHW6+jt06zrYjPx1n/vwDkx94SVZqLiEhAdOmTUNhYSHc3dmv7WzK31h5byUlJclKzQUEBGDnzp04dOiQrJCzKPu9Tp06WdvJaKz9h6nbtWuHfv36YcWKFbJCzrJ8+XKMHTsWeXl58PLiyH5XlrEvFft/+BxDfCbKCjnL2b9KsCH3HTzo9Rg0Hernk+uX2V1riIiIiIgaBwZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVcisV5POL5ubmhu7duyMnJ0dWyFmGDBmC9PR05Ofno0OHDrJaMz8Wf4fk7S+jVwcdbr6uo6ySs+zIT8eZP//A5AdekpWaW7rgacTGxmLVqlWyQs4UGhqKPn364JWEdbJSc7feeiuOHDki58jZlEPW2LFjsWzZMlmpuXbt2qFfv35YsWKFrJCzLF++3NpOeXl58PLyklVyRR/veAX5P++Dz61aWSFn2nloE3q2fxD33TtJVuqm3oK8soPs2bOnrJCzKDvFr776Crt27UK3bt1ktWY2bt2AxPdfQ1uG+AZxqqQYf/59Fu+/mSErNbduoQFeV5nkHDWEfcdux/DnvpZzNXf77bfjp59+wm233SYr5Ex79+696CB/3XXXoWPHjmyrBrJmzRp89NFH8PPzkxVyRTkbHkSLnzLlHDWE0x0nwueBt+Vc3dRbkKeGdbFBXqcdKueoIbi3a43C736WczWX+d543FyyDG26viwr5Ew/7Y/Dz6X3wnesWVZqLiAgADt37sShQ4dkhZxFuRLZqVMnJCUlwWg0ymrNXXnllfj777/lHDUEBnnXZ9n8GK4s3IKbOj8jK+RMP+5+Dv/zmgbv3m/ISt3UW5C/5557sGTJElkhZ3n66aeRmZl5UUH+06yt6NdbixUr30WP+3rIKjnLjKdm4uuvv8bXB76RlZr79asEHMueho4P/gdXNrleVslZ8j9+GM1vexTufZJkpeYY5BtOXYP8zTffjF69euG5556TFXKW7du348knn8TWrVvxwAMPyCq5op8+m4zfD61Ep4HslXe2v//8DQc3D8CNPeNwXedwWa0bjpFXmbqMkVfCf48ePbB27VoYDAZZJWfR6/WwWCzWqbYY5BsWg7w61DXIc4x8w+EYefVgkG84zgjyvGsNEREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERkYvLRoybG9zspxGpKJJLnc/28w0rq/mJO2LE72ZA6lE5T+RkDPJERETksopWGkQ49sXu1EKUlpbKqRApCIKHq4XmXpHid0tDYFs5T+RkDPJERETkmnbEwCPQhOjsUqQFuMuiwh2Bq0WYH25C0PSG7Jknci0M8kREROSCipA6LwoYngJjL1ly4A7t9CwUztPKeRtbD77dEJw52XKJ4twQmWykjrBbRw7TyZ5jV3OLEWuXZXYc4lN2eI/D0Brx+ys/Q/x8x9+poqsIcl27147ZIRcRVYFBnoiIiFxQAQrWiIfOniKyV8y9lwbubd3PL1eCuEcgkHLk3BCcLEQ/71sucJsCfVEwXa5zJAX6NcowHTfEdTw3fEfp7Y+Cb7ntguyG+MjhPdWN1Rc/PxSJdq9rQtAt9icJysmFh3ilFBRa1xFTdjSiNAzzVD0GeSIiInI9RwuwWzzoO3ra5qtzNBVxzwPR2fZj1DWIFKEYIqgn24fi2VmIPNfL31YLw3DxODwFieeH77hDO0wvthMnE7JiJba7MMTHHYHzbCcBDq9dVpnXDZwufh/xf1Yge+WLVsYhCnqkzAu8cMLSKxJZsyHCfEVXBYguYJAnIiIi9RPB3yQCsWfZD5r20sIanQsu9JvX+OSgjHLbyZOAqC1VxO0qrigoQ2rM60wi7BugLfN7awY6Bn6iijDIExERketp64mu4sGU79AnXqmiAqX/3rm6elYeyWvH5BjQ5dAe+zHybpoouZCocgzyRJelvYhp0x1uVU2v75Xr1pTtNQ3rf5Hzlal4vZJ8M2JChsND+dldhsMvehMsv8uFl70K7qFd4VQ/t+KzfTCP98KmS00D7WzxsK+g8jHo1g+X2saSu3sqsd+57Hv169XsLDmGvuzUCG5lWbgJhnPHlUqm6o8bZXyxRGw3E6mFcr5SFR9vineuRtioobaf7xuMCQtzUPSXXKgy6gnyR1NhKHfQqmiqj/Fk8tPj1X2AhSpU7o4BlU0OdxK4WDLg1MtrXYa0IxAX9WTF092t5UoN4MgmGB+biQV/9Ef8e4uQ/uJgNM2IgnfQEmT/Ide5rLnDJzYOceenSAR2EeUugYh0qOvg2dy2BV0iDXqsavwnXdbhJZWOQZd3tUE0tMp497ae0MOEtG1ljtw7zFDWqo/e9HJXB46akbZG/AYDNbJQW3Is/vPmcv8ebG3beMbI+zwWXPGxRky6W66SazlfiTgJ0D0ai903jBJttwhpEzxR8M5k+Ezf5Ph5CJVQT5Bv7gmdwwErFDqlPjjUrqZMPlWMRaOG0NRT59gmk60tBd1ku5oydWFLXXLdBiE8ZFTF08A2ciVnO4mMhVFIvSEEpndCEKjtIXaywUhLnQXdFwuxwFzLnppGSez/ZoQj/PwUBr9OotzJD2EOdSM0LW1b1IV7QFrj6Am8FHisql+9IlGYqrfewcXxW1WVDjcPBCkhOjsS1hjdNhARs5U7y4TandhkI0YZolLpLSxr6XlfuzvJiNe+JQimOr62e0CEOBUpc4cccUIYGmiCPlW8p2VJ7Ty1j1d8rBGTsVsLuZazfY3EZxYi2z8aptdHQN+7B/TBs5CeMApY+zpSv5SrqYh6gnxLDYwOB6wg+Cp1TZBdTZl04pBXV8oXTZSidLXdJ8ipxlr2Mjq2yRhrS8F3jF1Nmfzr3lLWOxIolx9nNZZdnYv66xeYl8TCT/uw9VKkh3YawpbsRbFcXLmzsKx/Q273MLxDViO3XA/7jygouAfa4b2haSJLiutaW9/LJX/bZqmmzt27OgPZc/3Q3s0N7UWQMZ+yLS05YEJM8AB4e8ie4U6+MMxMhUUuV5Tr5T13b+wDFqTO9JPbtodvcDyyj8t1yKZBj1WXx0mX9f/xSAoQ6GH7N2udLtyu8fzdZwTNrFIR/IGgW86tJ78Rtp6O5/pU8XuIk4pzrx2lDImp82srx7Fz31QrX1ucIHQt9yVYlwlnHm+OHUdhS0+EPtbboc2aXt8aHvhF/GxZUJFGOkbeNtzCsCIbpone4o3hAe8RybBYlxUje34YDJr28o0olmknIH67/T+RskNrzh0Ys1G8PR4T5LYenf0QsdKCEus6dFHOBYTtJkzoLHbSHt4wLLW1VM3aquzQGrZV/fsRqdMDMSDqADxGPY3MNW8i3r8JMqKC4TW16kuRue9MEjvTjSjpG4L0Nc8jqtMuBD272npLuQvuQOiqJciceIecF2Hz16+RGpuMxBsHIajn9bJKtbLQiKBtvogzpyM6xA9dlSE3+ckweBmQfEqLqCWZyDSnIcm/FbLmBkE7M6OaA2UBFoz3Q3KxDtGpmUhL0KLZxghxgp6sysvRrqGqY1UJLOtiMEGr1JX9n3JCZii3Hyt70nV+OMbxbMQH+1pP5JT9ql+ZkzXVaRuINIex45V3ttlObi6s5xiGbZ0/5b8ltvzr2V5H9vaf3y7Q1nl0birbidQrUtTPnVjJ161wHccTkPPr2r224/LLhZOPNzf2RtyaNVgw8MIVgJIjOYiJXYHcbiHQ3S2LKtKoP+yaNVOPuKbhSDYnIny0L7zE7s880wu+L1vQKjBOHMQykZ4qDkSnkxHRR4f4PbbtKpURDt34LHhOXYDMjBSEeVkQH6hF9FbGw7rJQvjwODSdnozMheEwarxEjW3lKoo3JiNobWtErn8LScFaaHv3RuBTsdj9nu1SZPJnZ+WaZfxqRuKLe6GZMRfpLz4Cndwuve9ZmOQq5dk+mNTs7iAELWmBuJTnofeQi6h2igxYkBoJfX8dAkdroIy4yd2SioJe0UhZHonAwaIt++thnJeGlGfdUbQoVwbIyuSiZGAK0t8JFa+phX5KElLm6YGNGcjmh2LrpPyxSpw2LTbAW5+ME/2jRF2cdJmSoL8hy7ofi9hQXd9kOsL9g5DlGYYF4kQuJcQLFuVk7WUzOzPIpTXo8UZ+CLdZz8mIKhiM9ORg+NhfFVaJRh3kizqLHWCCETrlYOUvdo3HzcjY1gz6eSlImqIXBzEtdAHhSFqRBB2ykWWp5qOtOzwQtiUNkQE6aAeLs/LVKYh2L0LM9ly5Al2cIng/n4wF48XfdZhorztFiW3VMOYG2z61X3Y6f8eak8j69AOg7wgY73P8MFJL7WBE3PULYrZ+WWE4KNm/F8noAeOwe9BU1hSe/sMRLp+X80cLaN5bhMzkWQgfeAARAych5otKdtxUteG+tl54Oz6TM5GnjCe2bxDROt6dlcEfJ3C6mpSne0jj0Ja2u4SUuY0e1Vq5Y5U4acpYUwDN7BSkzAoUdRFoxL4xTuzHIsV+LHFP1adcSk+/R0gm0qzbihO5WeJkbbY4WXslW7wyNWY7d+6Uz1yPKcQ2VKbsdOGOMg18vPnLE8Y1i5C+8EkYr30Pfg9FIfWIXKYijTrIo09Xa8/Gea11iMv+tswlNaGjt3UMY/Gp07b5ygwzQHObfG7lCc8+4qGqW2NRjWh9HFqKbdVQKrtrzfk71pxF8a/i4e42KN8x3hoe7cXDkV8qHJJR/OuP4r93wLPshk2ugcdd8nlZ//KEVttDnHw9grh3lyJJuxdRz6xm+KhnJcVFsOwQJ8srExEV4gftdKXPSrw3qhnv3uxK+YTqV9ljFXwQujkPWbMcT5zQ1Bveyn7s5Olqetb1MPRxHIHv2VHZc/LLhRq77Oxs3HfffUhPT5cV11HZXWsu3LGmgY83t9wBXW/l5gqjkLRqkQj8mxA0z1zNEEPX07iDfGX+KkHx4VyYt5qQPDcCEzRB1ltTmYuqiXjiINZMPqUGwrZyrsruWuPUO9a0QMsaDXtvA+2jWuCrvbBUe69gqpHDGYjQeqBZKw94ayYgeqkJRc11MI4Wf2dyQWL/V2RB9tYMpM6PQthQLcLXiPLhourDBk+6LksTJ07Ed999hyFDhqBbt2748MMP5ZJLr7K71jj3jjU1PN78qwcMQeLx/cPVDDF0PZdZkC9B7nwD2v+zGVp5dkXQ1ESk5pTAfaIRRrkGuQq2lWu4Ci2vEw/7f0T5LH0chd+Kh45tKvzAWcvrlJOBr1FQdsMzP8DyqXyuOPYZwrQPY8LGk7Jg54xSuwpNr7DNUl0UwzTbD/EngpB+5DRKS79F1uZMJMWGQq/cypJcSsGGCAzwEPs/D2/4ToxG8oYiNB1oRNBAuQJRBZo1a4aZM2dan+/evRuPPPKIywX6yjXA8ebL1fD1DUL8V3L+vLM4fUY83KgMNlSXyyvI70mEcaoJPgt348SfpSjcl4nM1QsQHWD7IBi5ELaVi2gB376PAJ+uRvJOx7HqxeaNiPvqeoR2q/gmek279EDojTlIXLPXYRhAwcYPkCKfW93YGp5X/ILkFLPjHQn+2IvUVTmAtgd8xM6V6soCy1Lx4G+Arq3doeovCzLWmeUMuYTjJsQMjceJ0ekoPF2K0oNZyNychLgp+jJDcIjKmzx5Mq677jrrnY4Uubm51kDfpUsXpKWlWWuuqQGON543o+vhrxG3LsdhPRwxI3kt4P54Z9W9xy6rIF9kyUKuMnZwsA9a2l12LN6a4djQdMmxrVxHy8FGpDx2HDGPTsKEJWaYP/sMqa/PRNdR7wGPPYVwbSWXRa/pjfAXB6Hw9Rnwm7Eaps9yYFo4EwNCNpX5nMIdCH01BBrzHAx4YglSzTkwb1yNCeNmIOrIIKS88ki93G+bvOAz2R2YEwTDXBPMW5Ux8jEI8tEi/iD/wi4l34Jk8aAbqoO7fffggQyYtsjnRJW4+uqr8fTTT1tvYak497hv3z489thjuPfee7F+/XprzdU4/XijrLdwELAw6vx6GeuXwPB4FFJvCYFpag/2yLsydy9faGBCeHAEkjeKfyBbTUhU7q8cbEKziq7V0CXDtnIlbRA4LxWZ0Xei8L3XMGD4NIRvOAP96ymwJAyqMmR7PhqN3DWj4XngPRiGT0bExhYIXR+LSLn8nKbdgpFhjoa+ZCPCR03GgGfXoLCjEZmboxF4i1yJ6qgldLFmpMzwgmWeAQO0QYhYXICusVnYvSYUPshC3sGqP0JJDaSjD0LFfi5mvAHx65T9XwZSxQmYd/94WDrKdVTi66+/xt69e2s1WSxqG6XsfMeOHavwb1XZ1LdvXzRv3vx8r7zif//7n/Vx//79GDZsGO655x6YTJXfDPjScP7xpux6frHb0Wp4LPJSgqH5l1xJRS6voTVdwpGxLQ5+p02YoBuAARPjYG6uR0puFhYME8uz8/jFJq6CbeVk9yDyx10ofeoeOV+NK6+HNngm0s3/QanYrtD8JuIev6PMMCfba6Y96vjJIvfeo5C04UPrdt9umIXw+7SIrmC9lncOQty7IsArv9eeNUh/ZQS0HFJTiaq+fbqKZc29EBibjrxC5QtnCpFnTkL4YE+0FO+33WI+ur+tL6rct4VW+AU2QmV1qhvlrl1bUxDuZUGcXuz/AiOQXNAVcdt2Iy3EB9iWhzyVnHMpgVHpAa7N9PDDD8utL4b8UkD5JVrlJ7tvLK6B81+wJecrZP1iQzfE7JDzTpCUlFTh36qySaPR4NSpU+d74+2dC/RKD73BYMCEWZvFetaSc3gMQloF+/xK1eR40y1YLItVvuzXQU2PN/brlWYtQdJTWnhdIxeqjJto5Do3n/IPuHv37sjJyZEVchblk+jKbaXy8/PRoUMHWa2ZXbt2oUePHli7dq31zUvOpdfrrT1LF9O79OtXCTiWPQ0dH/wPrmxSw50fXbT8jx9G89sehXufJFmpuYCAAOu9mw8dOiQr5CzKfq9Tp07WUGM01v5j7+3atUO/fv2wYsUKWSFnWb58OcaOHYu5c+fi5ptvltWaUYaGKGO6L44S5D0QhBQUVvLNr7WhBHmPwK7IOv8NrxVQgrwmCtHZzjup/eqrr6w97TWlRLvw8HD89NNP5cL8lVdeib/++gseHh548cUXMcRrN/4oWIVOAzPlGuQsf//5Gw5uHoAbe8bhus6V3uG+Vi6zu9YQERFRQ/Hz87Oe7NZmuvgQ33jdddddFf6tKpv+8Y9/oKioyCHEX3GF7fZfrVu3xvz583H48GHr7Sqv+MeF4TekPgzyREREdHmSw2LOTyNSq/3SQNtwG7uhOuXGeVYwvGdOlYNz6pUS3pWediXMK849tmrVCvPmzbMG+LCwMFx1leO3p5I6McgTERHRZccayOWQGCX8Kp8bSUEQPKoYE589xw0egUDKEbnNEQPSApWvKTzHbmiP9TWVKQvRz/vCsLK6U4T6odxiUvmQ8bmx8Ndeey1iY2OtXxT11FNPoUmTJtY6NQ4M8kRERNS4rFECeZlecYcPumYjOdAEfWqh3bh2dwTOS4EeUYirKHQfTUXc8xDbJF74IHjbQKRlR8sZRQEK1oh1hmntxudrECkCfVpAXUfsV085cXjhhResz1u0aIHo6Gh8//33mDFjhvWzB9T4MMgTERFR4zLcvkf83GR3J6ajBdgtHrp6lgnXbbUwDAdM+RXcF01sI6I/DH3KbuMpqudooJ0ttg/0ECcO1dztxgk++OADa8/7rFmzcOTIEURGRlpvQ0mNF4M8ERERXV6sobx2igqU6F8BEeS7yqcKzSxx0mDtpY+C7/mrAQ0T6ktKSqw98LNnz7YOqaHGj0GeiIiILi8Oveg14+5pH9ftyN59B/I7FqzTEdtwHd8GCPMjR47EddddJ+focsAgT0RERJcX2YsetaVMtD5qRpoyxr1jBd8hag3/JqRtKzN+vrre/baBSExVTht2o6AWX0ZFVBMM8kRERHSZ0cCohOvnfe2+kbUIqdODRCiPRkRFH0yVgdwUGOrwodkYjf1da8S8MpTG4XaTRTCvE1F/uAHac2P0ieoJgzwRERFddtwD0qxj2aM058axeyCocxZKq/gWV2WbwlQg6JZz2/gCqcrQmXOUO9TYbjd54W459fdNs0RlMcgTERFRI+GOwNWlKK1paLYfy65MsxwjvDXslwn2ttqFbSIDApGmPJ6/jaXtdpP269T49yGqJQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVyKxXk84vm5uaG7t27IycnR1bIWYYMGYL09HTk5+ejQ4cOslozu3btQo8ePbB27VoYDAZZJWfR6/WwWCzWqbZ+/SoBx7KnoeOD/8GVTa6XVXKW/I8fRvPbHoV7nyRZqblJI0di3759WLRokayQMyn7wDlz5mDstGmyUnPt2rVDv379sGLFClkhZ1m+fDnGjh2LvLw8eHl5ySq5op8+m4zfD61Ep4GZskLO8vefv+Hg5gG4sWccruscLqt1wyCvMgzy6lGXIP/6iyNx8tv3cc3Ng2WFnOnUT5+i6fVdEDF3u6zU3FsjDPAs/Z+co4bw6133IODFl+RczTHINxwGefX46q2xKPk5B81vekBWyJl+/2EjWng+hjvHzZWVuqm3IH/ttdeidevWskLOUlxcjF9++cUayrt16yarNaMcvMaMGVPrEwC6OH/++Sd+++03a5vV1qvPGfDsKyY5Rw0h0O92pHz0tZyruQ+fnIJr/zgFzz59ZYWcybLhQ/x2jw9GzHpeVmquefPmuOmmm6zHLHK+Q4cO4aOPPoKfn5+skCv6Yk4gju89LeeoIdwy0BPeE+fJubqptyCvTP/4B4fcO5vSXP/73//qFOSvuOIKWSFnU/7WZ86ckXM19/rcFzB95mx8s+8TWSFn6jPAAN0gLZa8+76s1NzeN+JRnPcVek4OlRVyltMnfsXnixbirkmhaDvgQVmtuSuvvFI+o4bw999/M8irwP6Fc3FsRy66jTfKCjnTjoUL0PHxoejw2HhZqRsOrVEZDq1Rj7oMrUlISMC0adNQWFgId3d3WSVnUf7GynsrKan2Y+QZ5BtOXYM8h9Y0HA6tUY+8xW+haNun6P1U/YzZpsr9dfo0tr8ejzvGjMNt/o/Iat2wC52IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnugylj3HDW5u1U+GlUVyizrYEWN9rZgdcp5q5qdNGNm1O5pWMY3c9ItcuYb2LRHbzcSqn+R8ZWqwXvGnsbi1Jq9FRJenS7kPw168Ws3rq30fpp4gfzQVhjLhouIpBtlyk4tXhNQR4rVGpIpnVFtFKw0VtEsF05y6t5SIooipt9e6/Lh3iUNc7IUpcrSPqPog8FnHus6zqW0DumS6+AXj1WlPVjgNuvkquVYD+3ETnpyzGszwdhr0WHVuf2tA6lFZIHJR3Ic5h3qCfHNP6OyCRVxsKHRKfXCoXU2ZfOBu3YAulaaeOsc2mWxtKegm29WUqQtb6lLz9A9H+IwLU9hgT6UKvxDHurFXS9sGddErEqWlpYjsJeepVm7r/TieHDuqwmlc5xZyrYb0I1a9GYWCNvfIebLisYqoQtyHOYd6gnxLDYx2wSJ8RhB8lbomyK6mTDoRQ+rKHYGrS1G6OpA72ovQspfRsU3GWFsKvmPsasrkX/eWEv8AECnCYeksjZwnZ7H1/MUgY0c8/Dq5wa2TL6K2lNgWHs9GYogBvkpd6W308MaA4HhkH7cttio3tMZ2NcWw0gLLygj4dfawLm+vmYD47cVyHaq1v3/B/62KxaMjHrZesr51xDQ8uWovqv2L/rdAHNSmocvA7mg6cDjGrfoasnUrcBafL47COETj3afulzWyatBjlThaBaSJE+Q0BLaVBSK1u9h9mNgvHdj0htzuYXR5bjX2/FcuKqfx7MMa6Rh5GRBWZMM00Rsebh7wHpEMi3VZMbLnh8GgaW8LHMoybdngUHZojZyfk43i7fGYILf16OyHCBFCKj/YUbWs4c6A1O0mTFCCnAiAhqW2lqpZW5UdWsO2cq4FMI42wzc2E+mzw+Dn0xQ4ZUZEZ19EH2iFIFHPNKcjZbovTi+NgK9/PHLllpUpSDDCb3ExdLNTkGlaAG2zdET0CUJyvlyBauFHrHoxEINiD8Bd/zQ2vf0mXhvYBJtig3HvrE04LNcq72u8MVmE94/OYtCUN7HpRSPuzI3CpNUH5HJHJfvew8zVbbBs2iDcdoUs0kWo6lhVAsu6GEzQKnVl/6ec5BrK7cfKDq05d8KdLU6u44N90V7ZVuxX/WamwnLKtg6R67rYfRiwJ2WSCO8bUdIzBOvffh7PeO7CuFdEmJfL7TWmfVij/rBr1kw94pqGI9mciPDRvvASuz/zTC/4vmxBq8A4EThEGEkVYeN0sggOOsRX1Nr2MsKhG58Fz6kLkJmRgjAvC+IDtYjeynhYN1kIHx6HptOTkbkwHEaNl6ixrVxTEQwJaYgcpoUuIBCa1uJ0a2sGsprrEb88CaGiru2vQ+CMJKQs0YkTtSxYqhm7m1uiRcqHctthoUhaHg89MpCRw0+o1Fbx1mSMS2+NZ5a8hbdGavFAj94Y+e9Y7Jg/Ckh/Hctyzso1HRVvXY1n9t2DF+e9iVeH9sYDvQfhmdfexKCzZrmGneLP8Mz09+Ez62mMbCNrVCflj1XiBHexAd76ZJzoHyXq4gTZlAT9DVnW/VjEhur6JtMR7h+ELM8wLFBOrEO8YJkbBO3LZnZmkEu72H0Yis14K34vek6ei/Xhj2CQ3G59r7NYL1c5r5Htwxp1kC/qLHaACUbo+uth9Be7xuNmZGxrBv28FCRN0YvAoYSRcCStSIIO2ciyVBMcdnggbIsIMQE6aAcHInJ1CqLdixCzvbo+R6paEbyfT8aC8eLvOky0152ixLZyUXr4dnb84GtL/zhkHSx/ad/TSxlQUIzT1SUHfz9omsvnirae6CoeTPkFtnmyWv+c7TJz2enC3RhOYseOD4BeIzCui+MHx1r2Hoyn7vgFr2Z9WUGQO4s9e5XtHsHIzvbbtcGjw8TB08GPWBU/G7kj5uLVvpdiTGvjVO5YhVxkrCmAZnYKUmYFirpykmtEnNiPRYr9WOKec1ctK5MNj5BMpFm3FSfWs9KQMtsdRa9kV3uFjMhZnLcPA0oO7MUy9MD4QffA/gh128DheFI+t2l8+7BGHeTRp6u1Z+O81jrEZX+LtIAyI987elvHMBafOm2br8wwAzS3yedWnvDsIx72FcghOHSxtD4OLcW2UqNTxSjYY4Z5XTLiZ06A7+goUTSjyH6cfEX+KR+pSpXd8eHC3R7OoljpqL2jTQWf7WkN93bi4cdf8JutYOckin4UDxVs17RVa3SRzxWHTa9h3PePIzbI8WBJdVT2WAUfhG7OQ9YsjePfuak3vJX92MnT1fSs62Ho4zgC37OjsufcjQLe3YYuEeftw4Dfiq07Mdx2k23+vKuugfsd8rnQGPdhjTvIV+avEhQfzoV5qwnJcyMwQRMEa+QoqibiXQk0k0+pgbCtXN+pXCSOaA+3Fq3Q3icIYfNTkXXKHcYxRrkC1YfK7vjg1Ls9XN0CF+5XtBeroj8TJ8ML0e9+ux61sQvFMjPGDVbml+Bz28pUL8T+r8iC7K0ZSJ0fhbChWoSvEeXDRdV/8E/sA4lcySXZh0Hsw1rJp410H3aZBfkS5M43oP0/m6GVZ1cETU1Eak4J3CcawcjhathWapG7yIiwNT5YkHsCpaWFyDNnIm1hNILq45aVVAtXoaXyJ//6xwquOh1H0ffi4bY2KNthpRzo3NuIhwq2K/mhAP8nnytXtR59exE2lZ2efUQs64FnXlPm+8Ou84vqoGBDBAZ4iP2fhzd8J0YjeUMRmg40ImigXIGo0bnYfRhwbUvrTgyHy94Q/uwPOHD+TmmNcx92eQX5PYkwTjXBZ+FunPizFIX7MpG5egGiAzR2vU7kEthWKlEES06udSiTrot9yxTDvCVFPqeG0QK9eokD0o7VWLbH8QNhxZ9txOtfX49J91R0w8Or0KX7CNxUbrsfsT59o3yuaIE7e/TAA2WnO2+2LbtHmffk+7M+HDchZmg8ToxOR+HpUpQezELm5iTETdGXGYJD1Jhc7D4MaHpXD0y6PgdvfbTXYdjZYfMHWCmfN9Z92GUV5IssWchVxg4O9kFLu8uOyl03GDlcC9tKLdzhpdEA68IxYXoyMrYqY+QTEabrCuO6ZvwehgbWsr8Ry/yO49XgSZi0yoz/y/kMqxbPRK8p7wF+T+HJ3hVfwm7ZexReU7abMU1uZ8YbT0/CuPRafm061Y98C5LFg26oDu72A3kPZMC0RT4naoQudh+GFr3xZPggFC6egUejV2N9Tg7WvzsTg57b1Oi/efqyCvLuXr7QwITw4AgkbxSBY6sJiSF+6BpsQjMmDpfCtlIPn6cykBXrh9MbJsBPOwATYs1oOjwFudsWwCCWZ1l495mG0wYjX0zFppl3osj0GgY9MQ1PbzmDR19MwZdzBsHh898O5Haj2sCyUhz8nngD65qNwCevBcvl1KA6+iBU7OdixhsQv07Z/2UgdU4QvPvHw9JRrkPUKF3sPgy4bVA0dr49GrcdfA8jn5iMZ8wtMGlJLJ6Ryxury2toTZdwZGyLg99pEyboBmDAxDiYm+uRkpuFBcPE8uw8MHK4CLbVJVHVt0RWvqwlNDOSkHWwVCwvxbfZaYgbr4G7uw4LxHz6eHkptFekdXlkL9ts5d/Ky2/rdXDTIKzavQurBl0vC9W44no8MHIm1q/+D0rEdt+tVu4Nf4fj5eLOwWJZLEbaDzZVths7C598sEss+xCfvDgKPQeGlF+vrIpei+pGuWvX1hSEe1kQpxf7v8AIJBd0Rdy23UgL8QG25SGvutu6ErmKhtqHCTf1GIW33v3Qut2Bd2fhyS5avFjdz1b5PsxNHFhL5fOLpnzjXPfu3ZGTkyMr5CxDhgxBeno68vPz0aFDB1mtmV27dqFHjx5Yu3YtDAalr5ScSa/Xw2KxWKfaSkhIwLRp01BYWCgCMS9BOJvyN1beW0lJSbJSc3vfiEdx3lfoOTlUVshZTp/4FZ8vWoi7JoWi7YAHZbXm2rVrh379+mHFihWyQs6yfPlyjB07Fnl5efDy4sh+V5a3+C0UbfsUvZ8KlxVylr9On8b21+Nxx5hxuM1f+ZBt3V1md60hIiIiImocGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFSIQZ6IiIiISIUY5ImIiIiIVIhBnoiIiIhIhRjkiYiIiIhUiEGeiIiIiEiFGOSJiIiIiFTIrVSQzy+am5sbunfvjpycHFkhZxkyZAjS09ORn5+PDh06yGrN7Nq1Cz169MDatWthMBhklZxFr9fDYrFYp9pKSEjAtGnTUFhYCHd3d1klZ1H+xsp7KykpSVZqLjooAP/94Qe07nS7rJAz/fhlLm7uPwBPvRYrKzXXrl079OvXDytWrJAVcpbly5dj7NixyMvLg5eXl6ySK8pb/BaKtn2K3k+Fywo5y1+nT2P76/G4Y8w43Ob/iKzWDYO8yjDIq0ddgrzSPiaTCXfccYeskDN9//331pB34MABWam5wV19sCl3j5yjhjA3fDoi4uLlXM0xyDccBnn1+CB6Of4oLETTli1lhZzpv78cR5t774R2cv3ksHoL8ldccYV1Iuf6+++/rZMSyrt16yarNaMcvMaMGYOrrrpKVqghnDlzRj6ruXNB/h//4Oi3htKxY0d8/fXXcq7mhj/6CL4QQT5t+buyQs7U9YH+SJw7FyEREbJSc69MXoB/uPE91VDO/nkWuuAH0E3TRVbIFSlB/th/m4r3hiyQU5098zfu6HCV6wX55s2b45prrpEVcpZTp07h999/r1OQb9OmjayQs508edLaXrXFoTUNqy5DawICArBz504cOnRIVshZlCuRnTp1sraT0WiU1Zp7c/pStGvfVs6RsxUdPo77DF4M8i7u07Rd+Cb3MB4KvF9WyFnOnvkTW1K2Q+Pvgy797pTVuuHQGpXh0Br14Bh59WCQV4e6BvkVcz7AtTe0QJe+3rJCznL0YBG+3GbByBl+aHUTO/lcGYN8w3FGkOc1RiIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJyIiIiJSIQZ5IiIiIiIVYpAnIiIiIlIhBnkiIiIiIhVikCciIiIiUiEGeSIiIiIiFWKQJ7qMZc9xg5tb9ZNhZZHcog52xFhfK2aHnKdayEZMmTapeDIg9ajcpA6KVhrq7bWI6DL3kwlP3OWBW6uYnsg4JleuoS/fFNsZ8cFPcr5SuzC/3Ovvx/x+9fR7uAAGeaLLmHuXOMTFXpgiR/uIqg8Cn3Ws6zyb2jagS8QdPnbtERcbicAuotwlEJEOdR08m9u2ICJyJXcNnYao8BcqnLQ3N+Ax5qdvsP840H/kJf496ol6gvzRVBgq7IEqO8UgW25y8YqQOkK81ohU8Yxqy9abV7ZdKpjm1L2lzvdU1strXX48/cMRPuPCFDbYU6nCL8SxbuzV0rZBXfSKRGlpKSJ7yXmqBU/o7NojfEYY/DqJcic/hDnUjdDUQ1O5B6SJtkpDYFtZoBr5OcNYYS/fhakmPYiObK/5JnbL+UpV1EP51ZvoXk+/B1Fd3dInGP8OnlTh9Pi918i1GsAP32ATdBh2qX+PeqKeIN9cHMgcep5CRTMIg0PtasrkA3frBnSpNPXUObbJZGtLQTfZrqZMXdhSamI7QYtBxo54ESLFyVMnX0RtKbEtPJ6NxBADfJW6cmLl4Y0BwfHIPm5bbFVuaI3tJMyw0gLLygj4dfawLm+vmYD47cVyHao92RExJwPZc/3Q3vo3jYL5lG1pyQETYoIHwNtDtpVoR8PMVFjkckW5oTXWthPzByxIneknt20P37JtTMLdeOTf5Xv6bNMA3HK1XK0B/PzdXvyMARhziX8Polr5+xiyUp7DuEfvtZ54dn90FGal7MLvcnHlziA/4yW53b0YMGMpvvqvXGTnyNGD4r/34JabbfNqp54g31IDo0PPUxB8lbomyK6mTDoofYp1447A1aUoXR3Ik4KL0LKX0bFNxlhbCr5j7GrK5F/3lhL/ABBZKtpqlkbOk3MtgHG0Gb6xmUifHQY/n6bAKTMiOvsi+kArBIl6pjkdKdN9cXppBHz945Ert6xMQYIRfouLoZudgkzTAmibpSOiTxCS8+UKdHEWGhG0zRdxoj2iQ/zQVRlyk58Mg5cByae0iFqitFUakvxbIWtuELQzM1D16VMBFoz3Q3KxDtGpmUhL0KLZRtHGY5LFErqgHQYElO/ps02B6NpCrtYAjhzKAB7UY+Il/j2Iau57fBA1EAEv78ONw1/ByqXvYdagJtj6sj+0z5pwRK5Vka+WPybCuwlnNE9j2dJ5mNphO6bOWYr9cvk5P38n3hdDb0CJEvoHK1eoemJY1FvIUmmnRCMdIy97+lZkwzTRGx5uHvAekQyLdVkxsueHwaBpb+uNUpZpy/YAlh1ac66HKxvF2+MxQW7r0dkPESstkH2SdDHO9fRtN2GC0iPr4Q3DUltL1aytyg6tYVs5VxEMCWmIHKaFLiAQmtailbZmIKu5HvHLkxAq6tr+OgTOSELKEp1o3yxYqvnAZG6JFikfym2HhSJpeTz0yEBGDge21UmRAQtSI6FX2mO0BsqIm9wtqSjoFY2U5ZEIHKy0lR7GeWlIedYdRYty5T6yMrkoGZiC9HdCxWtqoZ8i2nieHtiYgWx+KPbinDyITW+FY5g1THjg/sBwzP/4IM7IxZX670F8ED8KA5QP7PXri6kp+yvY5nscUU6GvdrhFluByOX9/vGbmPrhjZiyYi1ig3TwvW8AHpmUjIy3JgEfvoT3d1by7jiRgXdf24WuU5Kx7OlA9JfbLdOUYJNcxeZ7HNwrHj58DfM/b4eRL63BsrmBuPGLlxBgCMUHP9jWUpNG/WHXrJl6xDUNR7I5EeGjfeElYpx5phd8X7agVWAcMs2ZSE+NgO/pZET00SF+j227SmWEQzc+C55TFyAzIwVhXhbEB2oRvZXxsG6yED48Dk2nJyNzYTiMGi9RY1u5Jj18Ozt+GKilfxyyDpYfT+3ppVyJKcbp6v7k/n7Q2H9As60nuooHUz77eetkuK+tF96Oz+RM5GVHQuPQhE3h3VlpqxPVtpXuIY1Y+wJ3T2tLoYBBvvb+uwvzn+iLJ1b+hN5T3sPKpcmY2Ol7vDutL0a+tauKML8fiyeI8P7hGfR/UmwX/SQ6fRGCGauUdGLvGA5+DDzSusQa+u9XxsYP9sfMJdvx899yFaIGtGmGbahM2enCnWJ+xxfZqYDveDzetYms2VzTx4An7jyG+dtyKnxvnLHk4H3cj5G67rDf8paHx+Pf8rnV38fw+x/tbIH/pfEYdN/96K+bhrdXvyfWM2HqoowaDOFxLY06yBd1jkJyghE6pdfJX4TD42ZkbGsG/bwUJE3RQ9tf6VUMR9KKJOiQjSxLNT2AOzwQtiUNkQE6aAcHInJ1CqLdixCzvbrBA1S1Ing/n4wF48XfdZhorztFiW2lPqeKUbDHDPO6ZMTPnADf0VGiaEZRdZcr/ykfqcGUFBfBskO8x1YmIirED9rpJlEtqLatml0pn1CdfbXqWcR9qUNC6nuI0A2A7306jHlpDVbNGoDd89/A+u/limX8/vFSRH/ZHREJ7yFqmNiujx5T5r2HAWcy5BrS99/gC/Hwwfw38Zl7IGKXvoeER25EVvxwDI6qeogCkTNUdteaC3eKKcHvJ8TDne1wo61g50bceKt4+EEEcVvBwe/Fyhvmbtxyk23+vCYtcZOSKc65ojv+nfo51k1yDPxoMUBkDPFT1+Vgf7WXxFxLow7y6NMVSt/uea11iMv+FmkBZUa+d/S2jrcvPnXaNl+ZYQZobpPPrTzh2Uc87BMHQFuBLpLWx6Gl2FZqcioXiSPaw61FK7T3CULY/FRknXKHcYxRrkAu43AGIrQeaNbKA96aCYheakJRcx2Mo7VyBaq7DEwVf+PyPY/2d4rZj6z0/UDweAwq84G7jg8H4nFkYv3OipL8Gez/UumxDMCj99rHkHYY9Ngk+Vw68TtK2tkCf/khCqGY/7Ha+h1J7Sq7a41z7xRzDa65Tj6txg1tuov/fo+fVXavhcYd5CvzVwmKD+fCvNWE5LkRmKAJgrXvsKiaiHcl0Ew+pQbCtnJ5uYuMCFvjgwW5J1BaWog8cybSFkYjqD5uWUn1qBim2X6IPxGE9COnRVt9i6zNmUiKDYVeuZUl1ZPK7lpjf6eYEpQcEHn81naOvYKKVjdax7RnHa/oi2l+w8/KGN4KeiybXHcT7pLPre6dhHUbN2CKQ+BXhij4YWxr4P0v91Y/Fp+oQTXFNa3Ew4HvUf5f/zEc+048eLbDDbaCg2tathP/3Y8jZW+reuZ7HMySzxXW27R6IDqr7L/+MzhyMANofQ86lu3Vd3GXWZAvQe58A9r/sxlaeXZF0NREpOaUwH2iEew7dDVsK3UogiUn13oFRNfFPrgXw7wlRT4n12CBZal48Bdt1dZupPtfFmSsM8sZqrvK7lrj5DvFXH0NrpVPz6twLPyNaKd8mVglQxSILp1r0E0TKM5il+L93Y5B+/dtaXj7wI0Y0+V2WXHUpPP9GNN6O979wPHzJUcyV2K9fG51ZzdMESey6zekOw4v+yEd7y4B7ho7wPGEWAUuryC/JxHGqSb4LNyNE3+WonBfJjJXL0B0gO2ODuRC2FYq4Q4vjQZYF44J05ORsVUZI5+IMF1XGNc14+1bXYoXfCaLFpkTBMNcE8yirTJWxiDIR4v4g/VxK1iquaZoeqfIK999X75X/MQxa8AY1Kb8KGGIqH6DMhSngh7LM0cPwr7jcfdbHrj1npdQvuPxe+z7GLjB6/YKezaJLqVrHpyGhKHHMH/0Y5iZkoGsnZn44C0jdJPeAoa+gIl9KhmG02IAJj6txzGx7rgXlmLTzu3YtMSIgBkm/CxXsWpyPx4X6ynDywKmv2Vdb2vGqxgXGIpN9z6D6JF3yxXV47IK8kWWLORCD8NgH7S0+9CWcvs89h26FraVevg8JXa2sX44vWEC/LQDMCHWjKbDU5C7bQEMYnmWhXefcQ0toRNtkzLDC5Z5BgzQBiFicQG6xmZh95pQ+IgYmHeQd3VqGHfD108EhiUicJS53V3+f1LxPrqjt5cyVKCsJrj7vvG4oVyP5ffYtEH5wPIFd3WdJoK6CasyHcfaH/nPUiwWP/+JPuoLLHQ5aIdHordg5XOdcWzNswgYPwpzNp3BIFEzv6Kv8laqt+gSsXHpZNxy8C08MX44oj9uibErkjFFLj/Htt4L8C1aal1v3PzPcGNAMjKTpqGrCr8o7bIK8u5evtCIHVt4cASSN5qt464TlS9KCTahGbsOXQrb6tKo6qv5K1/WEpoZScg6WCqWl+Lb7DTEjdfA3V2HBWI+fbzs7e0VaV0e2cs2W/mXefFLvqpX1ZfWVbGsuRcCY9ORV6i0lfJ5hiSED/ZEyy7h2C3mo/vbhtyUa+tybSdVVqdq3TXyFUTcm4GpgaMQl5GJrJ0ZWP7CcIyck4muU17A45V8buGaPpMwS+mxfGqU7LHMwOLpwzH1Q8c++ib3BYr1gA9mDMcTS2TP5hujMOzZDPH6r2CM/Z08iJzpJj3e/qoQb+squspUgStuhG/Qy1i2/kt8J7bbtV65Q9PdcOiLv3eaWJaMR8qMZ7/hvkmITf3cut321Hj8u6sOERX8bPv1vtu4AbGTdOio0m87vryG1oiDVca2OPidNmGCbgAGTIyDubkeKblZWDBMLM/O4zcUugq2FRE1Zld3x5S3P8XbATfhs/mjEDDeiHcOtsMTb32OVWVvjedA9liObYeD7xnFdi8h4+rxWDdvmlx+jlwv/H4cW6OsNwpzd96EsW9+Ws3rE5GauJUq3Sl1pHxzZvfu3ZGTkyMr5CxDhgxBeno68vPz0aFDB1mtmV27dqFHjx5Yu3YtDAZl0AM5k16vh8VisU61lZCQgGnTpqGwsBDu7rwE4WzK31h5byUlJclKzQUEBGDnzp04dOiQrJCzKPu9Tp06WdvJaKz9x95XzPkA197QAl36essKOcvRg0X4cpsFI2f4odVNzry9INXVp2m78E3uYTwUeL+skLOcPfMntqRsh8bfB1361c9lscvsrjVERERERI0DgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQoxyBMRERERqRCDPBERERGRCjHIExERERGpEIM8EREREZEKMcgTEREREakQgzwRERERkQq5lQry+UVzc3ND9+7dkZOTIyvkLEOGDEF6ejry8/PRoUMHWa2ZXbt2oUePHli7di0MBoOskrPo9XpYLBbrVFsJCQmYNm0aCgsL4e7uLqvkLMrfWHlvJSUlyUrNBQQEYOfOnRg8eLCskDMlJiZa28loNMpKza2Y8wGuvaEFuvT1lhVylqMHi/DlNgtGzvBDq5uukVVyRZ+m7cI3uYfxUOD9skLOcvbMn9iSsh0afx906XenrNZNvQb59evXywo5yxNPPFEvQX7gwIGySs4ybty4Ogf5e+65R1bImX766ac6BfmMjAw5Rw1h3rx5dQryd/e6XVbIWYq+O84grxLngnyzq5vICjnT77+ecs0gTw2rLkGeGs6dd9550UF+8+bNco4agtIrfzFBntRDCfKnfvuvnKOGwCDv+pQgf6r4DzlHDaFNx5tcK8g//PDD8hk1lIULF15UkN+yZYuco4by7LPPymdEdCntzvxKPqOG4nn3LQzyRE5UL0GeiIiIiIgaFu9aQ0RERESkQgzyREREREQqxCBPRERERKRCDPJERERERCrEIE9EREREpEIM8kREREREKsQgT0RERESkQgzyREREREQqxCBPRERERKQ6wP8DJ1Mqy3pXVe0AAAAASUVORK5CYII="
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The method *fit_transform* internally relies on a cross fitting scheme to prevent target information from leaking into the train. Below is a diagram this process, but we will describe cross validation in more detail in the next section of this Notebook.\n",
    "\n",
    "<div>\n",
    "<img src=\"attachment:546e8541-dffc-45db-b3e6-05bfbbbd6f4e.png\" width=\"350\"/>\n",
    "</div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Multiple column transformations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What if we wanted to scale some numerical features, encode separately nominal and separately ordinal categorical features? So far, we have done column transformation steps separately.\n",
    "It would be more convenient to have a single transformer able to handle all columns, applying the appropriate transformations to each column. `sklearn` has `ColumnTransformer` for this purpose. Let's use it to apply all the transformations: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.compose import ColumnTransformer\n",
    "\n",
    "ct = ColumnTransformer([\n",
    "    ('scaling', MinMaxScaler(), ['age', 'hours-per-week']),\n",
    "    ('one_hot', OneHotEncoder(sparse_output=False, drop='first', \n",
    "                              min_frequency=0.01, handle_unknown='infrequent_if_exist' ), ['workclass']),\n",
    "    ('target_enc', TargetEncoder(target_type='binary'), ['occupation']),\n",
    "    ('ordinal', OrdinalEncoder(categories=categories), ['education','capital-gain-category'] )\n",
    "], remainder='passthrough')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The constructor of `ColumnTransfmer` requires a list of tuples, where each tuple contains a name of the transformation, user defined, a transformer and a list of names (or indices) of columns\n",
    "that the transformer should be applied to. In this example, we specify that the numerical\n",
    "columns should be transformed using the `MinMaxScaler`, that nominal feature workclass is encoded with `OneHotEncoder`, feature occupation will be encoded with `TargetEncoder` and the ordinal features with `OrdinalEncoder`. (We would also need to scale the ordinal encoded feature, but for now, we will ignore this step.)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we listed all the columns in the dataframe, as we wished to transform all of them. But we could have had a feature which we did not need to process additionally. By default, only the specified columns in transformers are transformed and combined in the output, and the non-specified columns are dropped. We can change the default behaviour, by specifying the parameter *remainder='passthrough'*, and with this, all remaining columns that were not specified in transformers, but present in the data passed to fit will be automatically passed through."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The default configuration for displaying a `ColumnTransformer` is 'text' where set_config(display='text'). To visualize the diagram in the notebook, use can use `set_config(display='diagram')` and then output the `ColumnTransformer` object."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-1 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-1 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-1 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-1 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-1 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-1 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-1 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-1 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-1 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-1 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>ColumnTransformer(remainder=&#x27;passthrough&#x27;,\n",
       "                  transformers=[(&#x27;scaling&#x27;, MinMaxScaler(),\n",
       "                                 [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                (&#x27;one_hot&#x27;,\n",
       "                                 OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                               handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                               min_frequency=0.01,\n",
       "                                               sparse_output=False),\n",
       "                                 [&#x27;workclass&#x27;]),\n",
       "                                (&#x27;target_enc&#x27;,\n",
       "                                 TargetEncoder(target_type=&#x27;binary&#x27;),\n",
       "                                 [&#x27;occupation&#x27;]),\n",
       "                                (&#x27;ordinal&#x27;,\n",
       "                                 OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                             &#x27;1st-4th&#x27;,\n",
       "                                                             &#x27;5th-6th&#x27;,\n",
       "                                                             &#x27;7th-8th&#x27;, &#x27;9th&#x27;,\n",
       "                                                             &#x27;10th&#x27;, &#x27;11th&#x27;,\n",
       "                                                             &#x27;12th&#x27;, &#x27;HS-grad&#x27;,\n",
       "                                                             &#x27;Prof-school&#x27;,\n",
       "                                                             &#x27;Assoc-acdm&#x27;,\n",
       "                                                             &#x27;Assoc-voc&#x27;,\n",
       "                                                             &#x27;Some-college&#x27;,\n",
       "                                                             &#x27;Bachelors&#x27;,\n",
       "                                                             &#x27;Masters&#x27;,\n",
       "                                                             &#x27;Doctorate&#x27;],\n",
       "                                                            [&#x27;cat1&#x27;, &#x27;cat2&#x27;,\n",
       "                                                             &#x27;cat3&#x27;, &#x27;cat4&#x27;]]),\n",
       "                                 [&#x27;education&#x27;, &#x27;capital-gain-category&#x27;])])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>ColumnTransformer</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for ColumnTransformer</span></a><span class=\"sk-estimator-doc-link \">i<span>Not fitted</span></span></div></label><div class=\"sk-toggleable__content \"><pre>ColumnTransformer(remainder=&#x27;passthrough&#x27;,\n",
       "                  transformers=[(&#x27;scaling&#x27;, MinMaxScaler(),\n",
       "                                 [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                (&#x27;one_hot&#x27;,\n",
       "                                 OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                               handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                               min_frequency=0.01,\n",
       "                                               sparse_output=False),\n",
       "                                 [&#x27;workclass&#x27;]),\n",
       "                                (&#x27;target_enc&#x27;,\n",
       "                                 TargetEncoder(target_type=&#x27;binary&#x27;),\n",
       "                                 [&#x27;occupation&#x27;]),\n",
       "                                (&#x27;ordinal&#x27;,\n",
       "                                 OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                             &#x27;1st-4th&#x27;,\n",
       "                                                             &#x27;5th-6th&#x27;,\n",
       "                                                             &#x27;7th-8th&#x27;, &#x27;9th&#x27;,\n",
       "                                                             &#x27;10th&#x27;, &#x27;11th&#x27;,\n",
       "                                                             &#x27;12th&#x27;, &#x27;HS-grad&#x27;,\n",
       "                                                             &#x27;Prof-school&#x27;,\n",
       "                                                             &#x27;Assoc-acdm&#x27;,\n",
       "                                                             &#x27;Assoc-voc&#x27;,\n",
       "                                                             &#x27;Some-college&#x27;,\n",
       "                                                             &#x27;Bachelors&#x27;,\n",
       "                                                             &#x27;Masters&#x27;,\n",
       "                                                             &#x27;Doctorate&#x27;],\n",
       "                                                            [&#x27;cat1&#x27;, &#x27;cat2&#x27;,\n",
       "                                                             &#x27;cat3&#x27;, &#x27;cat4&#x27;]]),\n",
       "                                 [&#x27;education&#x27;, &#x27;capital-gain-category&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>scaling</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;age&#x27;, &#x27;hours-per-week&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>MinMaxScaler</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.MinMaxScaler.html\">?<span>Documentation for MinMaxScaler</span></a></div></label><div class=\"sk-toggleable__content \"><pre>MinMaxScaler()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>one_hot</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;workclass&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-5\" type=\"checkbox\" ><label for=\"sk-estimator-id-5\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>OneHotEncoder</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></div></label><div class=\"sk-toggleable__content \"><pre>OneHotEncoder(drop=&#x27;first&#x27;, handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "              min_frequency=0.01, sparse_output=False)</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-6\" type=\"checkbox\" ><label for=\"sk-estimator-id-6\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>target_enc</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;occupation&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-7\" type=\"checkbox\" ><label for=\"sk-estimator-id-7\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>TargetEncoder</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.TargetEncoder.html\">?<span>Documentation for TargetEncoder</span></a></div></label><div class=\"sk-toggleable__content \"><pre>TargetEncoder(target_type=&#x27;binary&#x27;)</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-8\" type=\"checkbox\" ><label for=\"sk-estimator-id-8\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>ordinal</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;education&#x27;, &#x27;capital-gain-category&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-9\" type=\"checkbox\" ><label for=\"sk-estimator-id-9\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>OrdinalEncoder</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.OrdinalEncoder.html\">?<span>Documentation for OrdinalEncoder</span></a></div></label><div class=\"sk-toggleable__content \"><pre>OrdinalEncoder(categories=[[&#x27;Preschool&#x27;, &#x27;1st-4th&#x27;, &#x27;5th-6th&#x27;, &#x27;7th-8th&#x27;, &#x27;9th&#x27;,\n",
       "                            &#x27;10th&#x27;, &#x27;11th&#x27;, &#x27;12th&#x27;, &#x27;HS-grad&#x27;, &#x27;Prof-school&#x27;,\n",
       "                            &#x27;Assoc-acdm&#x27;, &#x27;Assoc-voc&#x27;, &#x27;Some-college&#x27;,\n",
       "                            &#x27;Bachelors&#x27;, &#x27;Masters&#x27;, &#x27;Doctorate&#x27;],\n",
       "                           [&#x27;cat1&#x27;, &#x27;cat2&#x27;, &#x27;cat3&#x27;, &#x27;cat4&#x27;]])</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-10\" type=\"checkbox\" ><label for=\"sk-estimator-id-10\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>remainder</div></div></label><div class=\"sk-toggleable__content \"><pre></pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-11\" type=\"checkbox\" ><label for=\"sk-estimator-id-11\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>passthrough</div></div></label><div class=\"sk-toggleable__content \"><pre>passthrough</pre></div> </div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "ColumnTransformer(remainder='passthrough',\n",
       "                  transformers=[('scaling', MinMaxScaler(),\n",
       "                                 ['age', 'hours-per-week']),\n",
       "                                ('one_hot',\n",
       "                                 OneHotEncoder(drop='first',\n",
       "                                               handle_unknown='infrequent_if_exist',\n",
       "                                               min_frequency=0.01,\n",
       "                                               sparse_output=False),\n",
       "                                 ['workclass']),\n",
       "                                ('target_enc',\n",
       "                                 TargetEncoder(target_type='binary'),\n",
       "                                 ['occupation']),\n",
       "                                ('ordinal',\n",
       "                                 OrdinalEncoder(categories=[['Preschool',\n",
       "                                                             '1st-4th',\n",
       "                                                             '5th-6th',\n",
       "                                                             '7th-8th', '9th',\n",
       "                                                             '10th', '11th',\n",
       "                                                             '12th', 'HS-grad',\n",
       "                                                             'Prof-school',\n",
       "                                                             'Assoc-acdm',\n",
       "                                                             'Assoc-voc',\n",
       "                                                             'Some-college',\n",
       "                                                             'Bachelors',\n",
       "                                                             'Masters',\n",
       "                                                             'Doctorate'],\n",
       "                                                            ['cat1', 'cat2',\n",
       "                                                             'cat3', 'cat4']]),\n",
       "                                 ['education', 'capital-gain-category'])])"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import set_config\n",
    "set_config(display=\"diagram\")\n",
    "ct"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's split our data into train and test. Then, let's fit our *ColumnTransformer* to our train data and then transform it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(24420, 12)"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)\n",
    "\n",
    "ct.fit(X_train, y_train)\n",
    "X_train_trans = ct.transform(X_train)\n",
    "X_train_trans.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that we obtained 12 features: 2 numerical, 2 ordinal and (9-2) dummy features for workclass, here we did not drop one dummy variable, as we will be using a regularized model later, but we grouped 2 infrequent categories together, and 1 target encoded feature, occupation. Let's see the list of the features for our transformed dataset, using *get_feature_names_out()* method on the fitted ColumnTransformer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['scaling__age', 'scaling__hours-per-week',\n",
       "       'one_hot__workclass_Federal-gov', 'one_hot__workclass_Local-gov',\n",
       "       'one_hot__workclass_Private', 'one_hot__workclass_Self-emp-inc',\n",
       "       'one_hot__workclass_Self-emp-not-inc',\n",
       "       'one_hot__workclass_State-gov',\n",
       "       'one_hot__workclass_infrequent_sklearn', 'target_enc__occupation',\n",
       "       'ordinal__education', 'ordinal__capital-gain-category'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ct.get_feature_names_out()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's access the one hot encoder inside the column transformer using the attribute `named_transformers_` to check that we did indeed end up with two categories grouped in *'one_hot__workclass_infrequent_sklearn'* category:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array(['Never-worked', 'Without-pay'], dtype=object)]"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ct.named_transformers_['one_hot'].infrequent_categories_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that we fitted the column transformer with X and y. Here we needed y only because we used traget encoding. If we did not use it, we would fit with only X."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we can build a *LogisticRegression* model. Since we transformed our train data, the same transformations must be done for the test data. As before, we only apply the *transform* on the test data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       <=50K       0.82      0.94      0.88      6159\n",
      "        >50K       0.66      0.38      0.48      1982\n",
      "\n",
      "    accuracy                           0.80      8141\n",
      "   macro avg       0.74      0.66      0.68      8141\n",
      "weighted avg       0.78      0.80      0.78      8141\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "logreg = LogisticRegression(solver=\"liblinear\", random_state=42)\n",
    "logreg.fit(X_train_trans, y_train)\n",
    "\n",
    "X_test_trans = ct.transform(X_test)\n",
    "y_pred=logreg.predict(X_test_trans)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cross validation and Hyperparameter tuning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To evaluate our supervised models, so far we have split our dataset into a training set and a test set using the `train_test_split` function, built a model on the training set by calling the `fit` method, and evaluated it on the test set using a variety of metrics.\n",
    "Note, the reason we split our data into training and test sets is that we are interested in measuring how well our model *generalizes* to new, previously unseen data. We are not interested in how well our model fit the training set, but rather in how\n",
    "well it can make predictions for data that was not observed during training."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cross-validation is a statistical method of evaluating generalization performance that is more stable and thorough than using a split into a training and a test set. In cross-validation, the data is instead split repeatedly and multiple models are trained. The\n",
    "most commonly used version of cross-validation is k-fold cross-validation, where k is a user-specified number, usually 5 or 10. When performing five-fold cross-validation, the data is first partitioned into five parts of (approximately) equal size, called folds.\n",
    "Next, a sequence of models is trained. The first model is trained using the first fold as the test set, and the remaining folds (2–5) are used as the training set. The model is built using the data in folds 2–5, and then the accuracy is evaluated on fold 1. Then\n",
    "another model is built, this time using fold 2 as the test set and the data in folds 1, 3, 4, and 5 as the training set. This process is repeated using folds 3, 4, and 5 as test sets. For each of these five splits of the data into training and test sets, we compute the accuracy. In the end, we have collected five accuracy values.  Usually, the first fifth of the data is the first fold, the second fifth of the data is the second fold, and so on. (Note: this is not valid for time series.)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's now import the iris dataset, using the parameter *return_X_y=True* to return the features and label directly. We will also shuffle the data, since the data is sorted by labels, and we want to have data instances of different labels in each fold. Later in this notebook, we will return to this and discuss better approach to achieve this, but for now, we will just shuffle the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "from sklearn.utils import shuffle\n",
    "X, y = load_iris(return_X_y=True)\n",
    "X, y = shuffle(X, y, random_state=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check the distribution of the target variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Counter({2: 50, 1: 50, 0: 50})"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from collections import Counter \n",
    "Counter(y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cross-validation is implemented in `sklearn` using the `cross_val_score` function from the `model_selection` module. The parameters of the `cross_val_score` function are the model we want to evaluate, the training data, and the ground-truth labels. By default, the score computed at each CV iteration is the score method of the estimator. For logistic regression this is accuracy. The following link gives a list of all the possible options that scoring parameter of `cross_val` can take: \\\n",
    "https://scikit-learn.org/stable/modules/model_evaluation.html  \\\n",
    "If we wanted to use another metric, we could pass another vale to `scoring` parameter, for example `scoring='f1'`. For now, let's keep the default option.\n",
    "\n",
    "Let’s evaluate LogisticRegression on the iris dataset: (we will not do scaling for now)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1.         0.86666667 0.96666667 0.96666667 0.96666667]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "log_reg = LogisticRegression(solver=\"liblinear\", random_state=42 , max_iter=10000)\n",
    "scores = cross_val_score(log_reg, X, y)\n",
    "print(scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default, `cross_val_score` performs five-fold cross-validation, returning five\n",
    "accuracy values. The number of folds can be changed using `cv` parameter."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A common way to summarize the cross-validation accuracy is to compute the mean:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average cross-validation score: 0.9533333333333334\n"
     ]
    }
   ],
   "source": [
    "print(\"Average cross-validation score:\", scores.mean() )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the mean cross-validation we can conclude that we expect the model to be around 95% accurate on average. Looking at all five scores produced by the five-fold cross-validation, we can also conclude that there is a relatively high variance in the\n",
    "accuracy between folds, ranging from 100% accuracy to 86% accuracy. This could imply that the model is very dependent on the particular folds used for training, but it could also just be a consequence of the small size of the dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that in cross-validation, for each fold, the process can be done in parallel, so for 5-fold CV, if we have computational resources, we could run 5 processes in parallel. With the parameter `n_jobs`, we could specify the number of jobs (number of concurrent threads or processes) to run in parallel. If this parameter is not specified, it means only 1 job is run at a time, while `n_jobs=-1` means that all available CPUs should be used. However, if the training takes a lot of time, allocating all CPUs to cross-validation will leave us without any resources to do other things on the computer.\n",
    "\n",
    "To check how many *logical* CPUs we have, we can use `os.cpu_count()`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "os.cpu_count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have used the default value of the hyperparameter C, which is inverse of regularization strength. However to try to get better model performance, we should try different values. Because hyperparameter tuning is such a common task, there are standard methods in `sklearn` that we can use. We can tune the parameters by using three sets: the training set to build the model, the validation  set to select the parameters of the model, and the test set to evaluate the performance of the selected parameters. \n",
    "After selecting the best parameters using the validation set, **we can rebuild a model using the parameter settings we found, but now training on both the training data and the validation data.** This way, we can use as much data as possible to build our model. The performance of the final model is evaluated on the test set."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Grid search"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The most commonly used method is grid search, which basically means trying all possible combinations of the\n",
    "parameters of interest."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_trainval, X_test, y_trainval, y_test = train_test_split(X, y, random_state=0, test_size=0.2)\n",
    "# split train+validation set into training and validation sets\n",
    "X_train, X_val, y_train, y_val = train_test_split(X_trainval, y_trainval, random_state=1, test_size=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Size of training set: 84, size of validation set: 36 size of test set 30\n"
     ]
    }
   ],
   "source": [
    "print(f'Size of training set: {X_train.shape[0]}, size of validation set: {X_val.shape[0]} size of test set {X_test.shape[0]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'C': 10, 'penalty': 'l1'}\n"
     ]
    }
   ],
   "source": [
    "best_score = 0\n",
    "for penalty_value in ['l1','l2']:\n",
    "    for C_value in [0.001, 0.01, 0.1, 1, 10]:\n",
    "        # for each combination of parameters, train a model\n",
    "        log_reg = LogisticRegression(solver=\"liblinear\", random_state=42, penalty=penalty_value, C=C_value, max_iter=10000)\n",
    "        log_reg.fit(X_train, y_train)\n",
    "        # evaluate the model on the validation set\n",
    "        score = log_reg.score(X_val, y_val)\n",
    "        # if we got a better score, store the score and parameters\n",
    "        if score > best_score:\n",
    "            best_score = score\n",
    "            best_parameters = {'C': C_value, 'penalty': penalty_value}\n",
    "print(best_parameters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best score on validation set 0.9722222222222222\n",
      "Test set score with best parameters 0.9666666666666667\n"
     ]
    }
   ],
   "source": [
    "# rebuild a model on the combined training and validation set,\n",
    "# and evaluate it on the test set\n",
    "log_reg = LogisticRegression(solver=\"liblinear\", random_state=42, penalty=best_parameters['penalty'], C=best_parameters['C'] , max_iter=10000)\n",
    "log_reg.fit(X_trainval, y_trainval)\n",
    "test_score = log_reg.score(X_test, y_test)\n",
    "print('Best score on validation set', best_score)\n",
    "print(\"Test set score with best parameters\", test_score)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The distinction between the training set, validation set, and test set is fundamentally important to applying machine learning methods in practice. Any choices made based on the test set accuracy \"leak\" information from the test set into the model.\n",
    "Therefore, it is important to keep a separate test set, which is only used for the final evaluation. It is good practice to do all exploratory analysis and model selection using the combination of a training and a validation set, and reserve the test set for a final evaluation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "While the method of splitting the data into a training, a validation, and a test set that\n",
    "we just saw is workable, and relatively commonly used, it is quite sensitive to how\n",
    "exactly the data is split. For a better estimate of the generalization performance, instead of\n",
    "using a single split into a training and a validation set, we can use cross-validation to\n",
    "evaluate the performance of each parameter combination. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Because grid search with cross-validation is such a commonly used method to adjust parameters, `sklearn` provides the `GridSearchCV` class, which implements it in the form of an estimator. To use the `GridSearchCV` class, we first need to specify the\n",
    "parameters we want to search over using a dictionary. `GridSearchCV` will then perform all the necessary model fits. The keys of the dictionary are the names of parameters we want to adjust (as given when constructing the model, in this case, *C* and\n",
    "*penalty*), and the values are the parameter settings we want to try out. The default metric for evaluation used will be the default score of the estimator, but just like in `cross_val_score`, a different `scoring` metric can be set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [],
   "source": [
    "param_grid = {'C': [0.001, 0.01, 0.1, 1, 10],\n",
    "               'penalty': ['l1','l2']}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Size of training set: 120, size of test set 30\n"
     ]
    }
   ],
   "source": [
    "print(f'Size of training set: {X_train.shape[0]}, size of test set {X_test.shape[0]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "grid_search = GridSearchCV(LogisticRegression(solver=\"liblinear\", random_state=42, max_iter=10000), param_grid, cv=5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The grid_search object that we created behaves just like a classifier; we can call the standard methods `fit`, `predict`, and `score` on it. However, when we call `fit`, it will run cross-validation for each combination of parameters we specified in the parameter grid. We have 5 different values for *C*, and 2 for *penalty*, giving in total 10 combinations. For each combination 5-fold CV is used, meaning in total 50 models are trained. And here we only have two parameters we want to tune. Hence, when running a grid search over many parameters and on large datasets it can be computationally challenging. One way to speed things up is with parallelization. Since using a particular parameter setting on a particular cross-validation split can be done completely independently from the other parameter settings and models, we can again parallelized just like we mentioned in `cross_val_score`. In GridSearchCV we could also use the parameter `n_jobs` to define how many jobs could be run in parallel."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-2 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-2 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-2 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-2 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-2 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-2 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-2 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-2 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-2 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-2 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-2 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-2 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-2 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-2 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-2 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-2 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>GridSearchCV(cv=5,\n",
       "             estimator=LogisticRegression(max_iter=10000, random_state=42,\n",
       "                                          solver=&#x27;liblinear&#x27;),\n",
       "             param_grid={&#x27;C&#x27;: [0.001, 0.01, 0.1, 1, 10],\n",
       "                         &#x27;penalty&#x27;: [&#x27;l1&#x27;, &#x27;l2&#x27;]})</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-12\" type=\"checkbox\" ><label for=\"sk-estimator-id-12\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>GridSearchCV</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.model_selection.GridSearchCV.html\">?<span>Documentation for GridSearchCV</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>GridSearchCV(cv=5,\n",
       "             estimator=LogisticRegression(max_iter=10000, random_state=42,\n",
       "                                          solver=&#x27;liblinear&#x27;),\n",
       "             param_grid={&#x27;C&#x27;: [0.001, 0.01, 0.1, 1, 10],\n",
       "                         &#x27;penalty&#x27;: [&#x27;l1&#x27;, &#x27;l2&#x27;]})</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-13\" type=\"checkbox\" ><label for=\"sk-estimator-id-13\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>best_estimator_: LogisticRegression</div></div></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(C=10, max_iter=10000, random_state=42, solver=&#x27;liblinear&#x27;)</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-14\" type=\"checkbox\" ><label for=\"sk-estimator-id-14\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>LogisticRegression</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.LogisticRegression.html\">?<span>Documentation for LogisticRegression</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(C=10, max_iter=10000, random_state=42, solver=&#x27;liblinear&#x27;)</pre></div> </div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "GridSearchCV(cv=5,\n",
       "             estimator=LogisticRegression(max_iter=10000, random_state=42,\n",
       "                                          solver='liblinear'),\n",
       "             param_grid={'C': [0.001, 0.01, 0.1, 1, 10],\n",
       "                         'penalty': ['l1', 'l2']})"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_search.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that this is fit, we can ask for the best parameters as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'C': 10, 'penalty': 'l2'}"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_search.best_params_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The best cross-validation accuracy (the mean accuracy over the different splits for this parameter setting) is stored in `best_score_`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.975"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_search.best_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The results of a grid search can be found\n",
    "in the `cv_results_` attribute, which is a dictionary storing all aspects of the search. We can convert it to a pandas dataframe to view it:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_C</th>\n",
       "      <th>param_penalty</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.002927</td>\n",
       "      <td>0.001664</td>\n",
       "      <td>0.003122</td>\n",
       "      <td>0.001338</td>\n",
       "      <td>0.001</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.001, 'penalty': 'l1'}</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.004115</td>\n",
       "      <td>0.002087</td>\n",
       "      <td>0.003950</td>\n",
       "      <td>0.001974</td>\n",
       "      <td>0.001</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 0.001, 'penalty': 'l2'}</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.325000</td>\n",
       "      <td>0.016667</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.003439</td>\n",
       "      <td>0.002083</td>\n",
       "      <td>0.001141</td>\n",
       "      <td>0.000825</td>\n",
       "      <td>0.010</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.01, 'penalty': 'l1'}</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.350000</td>\n",
       "      <td>0.020412</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   mean_fit_time  std_fit_time  mean_score_time  std_score_time  param_C  \\\n",
       "0       0.002927      0.001664         0.003122        0.001338    0.001   \n",
       "1       0.004115      0.002087         0.003950        0.001974    0.001   \n",
       "2       0.003439      0.002083         0.001141        0.000825    0.010   \n",
       "\n",
       "  param_penalty                         params  split0_test_score  \\\n",
       "0            l1  {'C': 0.001, 'penalty': 'l1'}           0.333333   \n",
       "1            l2  {'C': 0.001, 'penalty': 'l2'}           0.333333   \n",
       "2            l1   {'C': 0.01, 'penalty': 'l1'}           0.333333   \n",
       "\n",
       "   split1_test_score  split2_test_score  split3_test_score  split4_test_score  \\\n",
       "0           0.333333           0.333333           0.333333           0.333333   \n",
       "1           0.333333           0.333333           0.291667           0.333333   \n",
       "2           0.333333           0.333333           0.375000           0.375000   \n",
       "\n",
       "   mean_test_score  std_test_score  rank_test_score  \n",
       "0         0.333333        0.000000                9  \n",
       "1         0.325000        0.016667               10  \n",
       "2         0.350000        0.020412                8  "
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = pd.DataFrame(grid_search.cv_results_)\n",
    "results.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default an estimator is retrained using the best found parameters on the whole train dataset. We can access the model with `.best_estimator_` attribute and test its perfromance on the test set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9666666666666667"
      ]
     },
     "execution_count": 138,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_search.best_estimator_.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Or we could have just used the score method directly on the grid_search:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9666666666666667"
      ]
     },
     "execution_count": 140,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_search.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The important thing here is that **we did not use the test set to choose the parameters**, meaning the final test set was only used for the final model evaluation. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Randomized search"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we had many different paramenters and many values to evaluate, randomized search would be a better option than grid. In contrast to `GridSearchCV`, not all parameter values are tried out, but rather a fixed number of parameter settings is sampled from the specified distributions. The number of parameter settings that are tried is given by *n_iter*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.stats import loguniform\n",
    "\n",
    "distributions= { 'C': loguniform(0.0001, 10), \n",
    "                 'penalty': ['l1', 'l2']}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using a loguniform distribution instead of a regular uniform distribution will ensure that in a sufficiently large number of trials, the same number of samples will be drawn from the [0.0001, 0.001] range as, for example, the [1, 10] range."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'C': 4.600306804490298, 'penalty': 'l2'}"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "random_search = RandomizedSearchCV(LogisticRegression(solver=\"liblinear\", random_state=42, max_iter=10000), distributions, cv=5,  n_iter=10, random_state=1)\n",
    "random_search.fit(X_train, y_train)\n",
    "random_search.best_params_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Running the results with a different seed especially for a small number of iterations might lead to different parameters selected. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we evaluate the performance of the model with the best parameters on the test set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9666666666666667"
      ]
     },
     "execution_count": 150,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "random_search.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Just like in `GridSearchCV` we could change the `scoring` method, use parallelization with `n_jobs`, view `cv_results_` and access the model through `best_estimator`.\n",
    "\n",
    "If we had a large number of parameters,  would me much more practical choice than `GridSearchCV`. However, there are more efficient approached than `RandomizedSearchCV`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Successive Halving search"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Sklearn* also provides the *HalvingGridSearchCV* and *HalvingRandomSearchCV* estimators that can be used to search a parameter space using successive halving . Successive halving (SH) is like a tournament among candidate hyperparameter combinations. It is an iterative selection process where all candidates (the hyperparameter combinations) are evaluated with a small amount of resources at the first iteration. Only some of these candidates are selected for the next iteration, which will be allocated more resources. For parameter tuning, the resource is typically the number of training samples, but it can also be an arbitrary numeric parameter."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can summarize the procedure via the following steps:\n",
    "\n",
    "1. Draw a large set of candidate configurations (hyperparameter combinations) via random sampling.\n",
    "2. Train the models (i.e., each set of hyperparameter combinations) with limited resources, for example, a small subset of the training data (as opposed to using the entire training set).\n",
    "3. Discard the bottom 50 percent (typical value) based on predictive performance.\n",
    "4. Go back to step 2 with an increased amount of available resources for each of the surviving configurations.\n",
    "\n",
    "The steps are repeated until only one hyperparameter configuration remains."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.experimental import enable_halving_search_cv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'C': 4.600306804490298, 'penalty': 'l2'}"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import HalvingRandomSearchCV\n",
    "hs = HalvingRandomSearchCV(LogisticRegression(solver=\"liblinear\", random_state=42, max_iter=10000), \n",
    "                            param_distributions=distributions, \n",
    "# specifies the resource that will be allocated to each candidate configuration during evaluation. \n",
    "# In this case, it's set to 'n_samples', indicating that the number of training samples will be used as the resource. \n",
    "# Each candidate configuration will be trained and evaluated using a subset of the training data.\n",
    "                            resource='n_samples', \n",
    "# select half of candidates in each iteration % (default is 3, select 1/3 of candidates)                         \n",
    "                           factor=2, \n",
    "# The number of candidate parameters to sample, at the first iteration.\n",
    "                           n_candidates=10,\n",
    "                           random_state=1)\n",
    "\n",
    "hs.fit(X_train, y_train)\n",
    "hs.best_params_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Again, we evaluate the performance of the model with the best parameters on the test set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9666666666666667"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hs.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `cv_results_` attribute again contains useful information for analyzing the results of a search."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>iter</th>\n",
       "      <th>n_resources</th>\n",
       "      <th>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_C</th>\n",
       "      <th>param_penalty</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "      <th>split0_train_score</th>\n",
       "      <th>split1_train_score</th>\n",
       "      <th>split2_train_score</th>\n",
       "      <th>split3_train_score</th>\n",
       "      <th>split4_train_score</th>\n",
       "      <th>mean_train_score</th>\n",
       "      <th>std_train_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.004911</td>\n",
       "      <td>0.001739</td>\n",
       "      <td>0.003597</td>\n",
       "      <td>0.002706</td>\n",
       "      <td>0.012165</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.012164941464151846, 'penalty': 'l1'}</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.433333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>7</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.308333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.009551</td>\n",
       "      <td>0.006414</td>\n",
       "      <td>0.008887</td>\n",
       "      <td>0.011955</td>\n",
       "      <td>4.600307</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 4.600306804490298, 'penalty': 'l2'}</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.866667</td>\n",
       "      <td>0.124722</td>\n",
       "      <td>4</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>0.875000</td>\n",
       "      <td>0.875000</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>0.037268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.011571</td>\n",
       "      <td>0.016472</td>\n",
       "      <td>0.004193</td>\n",
       "      <td>0.004823</td>\n",
       "      <td>0.003248</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 0.003248350345086679, 'penalty': 'l2'}</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.233333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>15</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.541667</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.483333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.006710</td>\n",
       "      <td>0.006244</td>\n",
       "      <td>0.001695</td>\n",
       "      <td>0.001438</td>\n",
       "      <td>0.001515</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 0.0015151125123102904, 'penalty': 'l2'}</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.233333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>15</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.541667</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.483333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.007452</td>\n",
       "      <td>0.003027</td>\n",
       "      <td>0.012958</td>\n",
       "      <td>0.010096</td>\n",
       "      <td>0.000854</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 0.0008536916958038761, 'penalty': 'l2'}</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.233333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>15</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.541667</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.483333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.003607</td>\n",
       "      <td>0.003976</td>\n",
       "      <td>0.005699</td>\n",
       "      <td>0.002087</td>\n",
       "      <td>0.223218</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 0.223218499287176, 'penalty': 'l2'}</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.105409</td>\n",
       "      <td>6</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.026352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.004812</td>\n",
       "      <td>0.003195</td>\n",
       "      <td>0.000630</td>\n",
       "      <td>0.001260</td>\n",
       "      <td>0.049441</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.04944059287398676, 'penalty': 'l1'}</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.233333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>15</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.541667</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.483333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.004658</td>\n",
       "      <td>0.001181</td>\n",
       "      <td>0.005618</td>\n",
       "      <td>0.004366</td>\n",
       "      <td>0.003684</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.0036844068804921994, 'penalty': 'l1'}</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.433333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>7</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.308333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.004179</td>\n",
       "      <td>0.003769</td>\n",
       "      <td>0.004323</td>\n",
       "      <td>0.006101</td>\n",
       "      <td>0.001053</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.001052594868979971, 'penalty': 'l1'}</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.433333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>7</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.308333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>0.008003</td>\n",
       "      <td>0.005219</td>\n",
       "      <td>0.002800</td>\n",
       "      <td>0.003768</td>\n",
       "      <td>0.001406</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.0014056787147388348, 'penalty': 'l1'}</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.433333</td>\n",
       "      <td>0.081650</td>\n",
       "      <td>7</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.308333</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>0.005331</td>\n",
       "      <td>0.007196</td>\n",
       "      <td>0.009976</td>\n",
       "      <td>0.009796</td>\n",
       "      <td>0.003684</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.0036844068804921994, 'penalty': 'l1'}</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.416667</td>\n",
       "      <td>0.350000</td>\n",
       "      <td>0.033333</td>\n",
       "      <td>11</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.320833</td>\n",
       "      <td>0.028260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>0.006290</td>\n",
       "      <td>0.005508</td>\n",
       "      <td>0.004830</td>\n",
       "      <td>0.005275</td>\n",
       "      <td>0.001053</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.001052594868979971, 'penalty': 'l1'}</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.416667</td>\n",
       "      <td>0.350000</td>\n",
       "      <td>0.033333</td>\n",
       "      <td>11</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.320833</td>\n",
       "      <td>0.028260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>0.006185</td>\n",
       "      <td>0.004579</td>\n",
       "      <td>0.003850</td>\n",
       "      <td>0.004312</td>\n",
       "      <td>0.001406</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.0014056787147388348, 'penalty': 'l1'}</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.416667</td>\n",
       "      <td>0.350000</td>\n",
       "      <td>0.033333</td>\n",
       "      <td>11</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.291667</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>0.320833</td>\n",
       "      <td>0.028260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>0.005531</td>\n",
       "      <td>0.005763</td>\n",
       "      <td>0.007023</td>\n",
       "      <td>0.005831</td>\n",
       "      <td>0.223218</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 0.223218499287176, 'penalty': 'l2'}</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.783333</td>\n",
       "      <td>0.124722</td>\n",
       "      <td>5</td>\n",
       "      <td>0.812500</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.791667</td>\n",
       "      <td>0.937500</td>\n",
       "      <td>0.854167</td>\n",
       "      <td>0.837500</td>\n",
       "      <td>0.054962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>0.007699</td>\n",
       "      <td>0.002374</td>\n",
       "      <td>0.003544</td>\n",
       "      <td>0.001680</td>\n",
       "      <td>4.600307</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 4.600306804490298, 'penalty': 'l2'}</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.933333</td>\n",
       "      <td>0.062361</td>\n",
       "      <td>2</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.979167</td>\n",
       "      <td>0.970833</td>\n",
       "      <td>0.016667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>2</td>\n",
       "      <td>120</td>\n",
       "      <td>0.007281</td>\n",
       "      <td>0.002673</td>\n",
       "      <td>0.007458</td>\n",
       "      <td>0.003215</td>\n",
       "      <td>0.001406</td>\n",
       "      <td>l1</td>\n",
       "      <td>{'C': 0.0014056787147388348, 'penalty': 'l1'}</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>14</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>2</td>\n",
       "      <td>120</td>\n",
       "      <td>0.005487</td>\n",
       "      <td>0.003290</td>\n",
       "      <td>0.006534</td>\n",
       "      <td>0.005949</td>\n",
       "      <td>0.223218</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 0.223218499287176, 'penalty': 'l2'}</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.875000</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.908333</td>\n",
       "      <td>0.048591</td>\n",
       "      <td>3</td>\n",
       "      <td>0.947917</td>\n",
       "      <td>0.927083</td>\n",
       "      <td>0.937500</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>0.895833</td>\n",
       "      <td>0.925000</td>\n",
       "      <td>0.017922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>2</td>\n",
       "      <td>120</td>\n",
       "      <td>0.006314</td>\n",
       "      <td>0.004630</td>\n",
       "      <td>0.002907</td>\n",
       "      <td>0.001485</td>\n",
       "      <td>4.600307</td>\n",
       "      <td>l2</td>\n",
       "      <td>{'C': 4.600306804490298, 'penalty': 'l2'}</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.966667</td>\n",
       "      <td>0.016667</td>\n",
       "      <td>1</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.968750</td>\n",
       "      <td>0.968750</td>\n",
       "      <td>0.979167</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.966667</td>\n",
       "      <td>0.007795</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    iter  n_resources  mean_fit_time  std_fit_time  mean_score_time  \\\n",
       "0      0           30       0.004911      0.001739         0.003597   \n",
       "1      0           30       0.009551      0.006414         0.008887   \n",
       "2      0           30       0.011571      0.016472         0.004193   \n",
       "3      0           30       0.006710      0.006244         0.001695   \n",
       "4      0           30       0.007452      0.003027         0.012958   \n",
       "5      0           30       0.003607      0.003976         0.005699   \n",
       "6      0           30       0.004812      0.003195         0.000630   \n",
       "7      0           30       0.004658      0.001181         0.005618   \n",
       "8      0           30       0.004179      0.003769         0.004323   \n",
       "9      0           30       0.008003      0.005219         0.002800   \n",
       "10     1           60       0.005331      0.007196         0.009976   \n",
       "11     1           60       0.006290      0.005508         0.004830   \n",
       "12     1           60       0.006185      0.004579         0.003850   \n",
       "13     1           60       0.005531      0.005763         0.007023   \n",
       "14     1           60       0.007699      0.002374         0.003544   \n",
       "15     2          120       0.007281      0.002673         0.007458   \n",
       "16     2          120       0.005487      0.003290         0.006534   \n",
       "17     2          120       0.006314      0.004630         0.002907   \n",
       "\n",
       "    std_score_time   param_C param_penalty  \\\n",
       "0         0.002706  0.012165            l1   \n",
       "1         0.011955  4.600307            l2   \n",
       "2         0.004823  0.003248            l2   \n",
       "3         0.001438  0.001515            l2   \n",
       "4         0.010096  0.000854            l2   \n",
       "5         0.002087  0.223218            l2   \n",
       "6         0.001260  0.049441            l1   \n",
       "7         0.004366  0.003684            l1   \n",
       "8         0.006101  0.001053            l1   \n",
       "9         0.003768  0.001406            l1   \n",
       "10        0.009796  0.003684            l1   \n",
       "11        0.005275  0.001053            l1   \n",
       "12        0.004312  0.001406            l1   \n",
       "13        0.005831  0.223218            l2   \n",
       "14        0.001680  4.600307            l2   \n",
       "15        0.003215  0.001406            l1   \n",
       "16        0.005949  0.223218            l2   \n",
       "17        0.001485  4.600307            l2   \n",
       "\n",
       "                                           params  split0_test_score  \\\n",
       "0    {'C': 0.012164941464151846, 'penalty': 'l1'}           0.500000   \n",
       "1       {'C': 4.600306804490298, 'penalty': 'l2'}           0.833333   \n",
       "2    {'C': 0.003248350345086679, 'penalty': 'l2'}           0.166667   \n",
       "3   {'C': 0.0015151125123102904, 'penalty': 'l2'}           0.166667   \n",
       "4   {'C': 0.0008536916958038761, 'penalty': 'l2'}           0.166667   \n",
       "5       {'C': 0.223218499287176, 'penalty': 'l2'}           0.666667   \n",
       "6     {'C': 0.04944059287398676, 'penalty': 'l1'}           0.166667   \n",
       "7   {'C': 0.0036844068804921994, 'penalty': 'l1'}           0.500000   \n",
       "8    {'C': 0.001052594868979971, 'penalty': 'l1'}           0.500000   \n",
       "9   {'C': 0.0014056787147388348, 'penalty': 'l1'}           0.500000   \n",
       "10  {'C': 0.0036844068804921994, 'penalty': 'l1'}           0.333333   \n",
       "11   {'C': 0.001052594868979971, 'penalty': 'l1'}           0.333333   \n",
       "12  {'C': 0.0014056787147388348, 'penalty': 'l1'}           0.333333   \n",
       "13      {'C': 0.223218499287176, 'penalty': 'l2'}           0.750000   \n",
       "14      {'C': 4.600306804490298, 'penalty': 'l2'}           0.916667   \n",
       "15  {'C': 0.0014056787147388348, 'penalty': 'l1'}           0.333333   \n",
       "16      {'C': 0.223218499287176, 'penalty': 'l2'}           0.833333   \n",
       "17      {'C': 4.600306804490298, 'penalty': 'l2'}           0.958333   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "0            0.500000           0.333333           0.333333   \n",
       "1            1.000000           0.833333           0.666667   \n",
       "2            0.166667           0.333333           0.166667   \n",
       "3            0.166667           0.333333           0.166667   \n",
       "4            0.166667           0.333333           0.166667   \n",
       "5            0.666667           0.666667           0.500000   \n",
       "6            0.166667           0.333333           0.166667   \n",
       "7            0.500000           0.333333           0.333333   \n",
       "8            0.500000           0.333333           0.333333   \n",
       "9            0.500000           0.333333           0.333333   \n",
       "10           0.333333           0.333333           0.333333   \n",
       "11           0.333333           0.333333           0.333333   \n",
       "12           0.333333           0.333333           0.333333   \n",
       "13           0.666667           0.666667           0.833333   \n",
       "14           1.000000           0.833333           0.916667   \n",
       "15           0.333333           0.333333           0.333333   \n",
       "16           0.916667           0.958333           0.875000   \n",
       "17           1.000000           0.958333           0.958333   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \\\n",
       "0            0.500000         0.433333        0.081650                7   \n",
       "1            1.000000         0.866667        0.124722                4   \n",
       "2            0.333333         0.233333        0.081650               15   \n",
       "3            0.333333         0.233333        0.081650               15   \n",
       "4            0.333333         0.233333        0.081650               15   \n",
       "5            0.833333         0.666667        0.105409                6   \n",
       "6            0.333333         0.233333        0.081650               15   \n",
       "7            0.500000         0.433333        0.081650                7   \n",
       "8            0.500000         0.433333        0.081650                7   \n",
       "9            0.500000         0.433333        0.081650                7   \n",
       "10           0.416667         0.350000        0.033333               11   \n",
       "11           0.416667         0.350000        0.033333               11   \n",
       "12           0.416667         0.350000        0.033333               11   \n",
       "13           1.000000         0.783333        0.124722                5   \n",
       "14           1.000000         0.933333        0.062361                2   \n",
       "15           0.333333         0.333333        0.000000               14   \n",
       "16           0.958333         0.908333        0.048591                3   \n",
       "17           0.958333         0.966667        0.016667                1   \n",
       "\n",
       "    split0_train_score  split1_train_score  split2_train_score  \\\n",
       "0             0.333333            0.250000            0.333333   \n",
       "1             0.916667            0.875000            0.875000   \n",
       "2             0.500000            0.541667            0.458333   \n",
       "3             0.500000            0.541667            0.458333   \n",
       "4             0.500000            0.541667            0.458333   \n",
       "5             0.833333            0.791667            0.791667   \n",
       "6             0.500000            0.541667            0.458333   \n",
       "7             0.333333            0.250000            0.333333   \n",
       "8             0.333333            0.250000            0.333333   \n",
       "9             0.333333            0.250000            0.333333   \n",
       "10            0.312500            0.291667            0.375000   \n",
       "11            0.312500            0.291667            0.375000   \n",
       "12            0.312500            0.291667            0.375000   \n",
       "13            0.812500            0.791667            0.791667   \n",
       "14            0.958333            0.958333            0.958333   \n",
       "15            0.333333            0.333333            0.333333   \n",
       "16            0.947917            0.927083            0.937500   \n",
       "17            0.958333            0.968750            0.968750   \n",
       "\n",
       "    split3_train_score  split4_train_score  mean_train_score  std_train_score  \n",
       "0             0.333333            0.291667          0.308333         0.033333  \n",
       "1             0.958333            0.958333          0.916667         0.037268  \n",
       "2             0.458333            0.458333          0.483333         0.033333  \n",
       "3             0.458333            0.458333          0.483333         0.033333  \n",
       "4             0.458333            0.458333          0.483333         0.033333  \n",
       "5             0.791667            0.750000          0.791667         0.026352  \n",
       "6             0.458333            0.458333          0.483333         0.033333  \n",
       "7             0.333333            0.291667          0.308333         0.033333  \n",
       "8             0.333333            0.291667          0.308333         0.033333  \n",
       "9             0.333333            0.291667          0.308333         0.033333  \n",
       "10            0.312500            0.312500          0.320833         0.028260  \n",
       "11            0.312500            0.312500          0.320833         0.028260  \n",
       "12            0.312500            0.312500          0.320833         0.028260  \n",
       "13            0.937500            0.854167          0.837500         0.054962  \n",
       "14            1.000000            0.979167          0.970833         0.016667  \n",
       "15            0.333333            0.333333          0.333333         0.000000  \n",
       "16            0.916667            0.895833          0.925000         0.017922  \n",
       "17            0.979167            0.958333          0.966667         0.007795  "
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = pd.DataFrame(hs.cv_results_)\n",
    "results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the results we can see that in the first round (iter 0), 10 different hyperparameter combinations were used. Since a factor of 2 was specified, this means only 5 combinations were selected for the next round with double the number of resources, i.e., the number of training data points used increased from 30 to 60. Then, half of these 5 combinations were selected, in our case, this is 3, to go to the next round, and the models were trained on all the samples. Out of those 3 combinations, the one with the highest *mean_test_score* (lowest `rank_test_score`) was selected as the best one."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Other libraries for hyperparameter tuning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since hyperparameter tuning is a very important step in modeling, many different approaches and libraries were developed to speed up the model tuning when there are many different parameters to be tuned. Here, we will illustrate one such simple library, `scikit-optimize`, which implements Bayesian optimization, which works by considering the previously seen hyperparameter combinations when determining the next set of hyperparameters to evaluate. Bayesian optimization reduces the number of evaluations needed to identify optimal hyperparameters, making it effective for optimizing complex models.\n",
    "Let's install the `scikit-optimize` library with:\n",
    "\n",
    "`pip install scikit-optimize`\n",
    "\n",
    "Note that when importing this library is referred to as: `skopt`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [],
   "source": [
    "from skopt import BayesSearchCV\n",
    "from skopt.space import Real "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [],
   "source": [
    "distributions = {\n",
    "    'penalty': ['l1','l2'],\n",
    "    'C': Real(low=1e-4, high=10,  prior='log-uniform'),\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-3 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-3 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-3 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-3 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-3 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-3 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-3 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-3 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-3 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-3 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-3 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-3 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-3 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-3 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-3 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-3 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-3 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-3 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-3 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-3\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>BayesSearchCV(estimator=LogisticRegression(max_iter=10000, random_state=42,\n",
       "                                           solver=&#x27;liblinear&#x27;),\n",
       "              n_iter=10, random_state=1,\n",
       "              search_spaces={&#x27;C&#x27;: Real(low=0.0001, high=10, prior=&#x27;log-uniform&#x27;, transform=&#x27;normalize&#x27;),\n",
       "                             &#x27;penalty&#x27;: [&#x27;l1&#x27;, &#x27;l2&#x27;]})</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-15\" type=\"checkbox\" ><label for=\"sk-estimator-id-15\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>BayesSearchCV</div></div><div><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>BayesSearchCV(estimator=LogisticRegression(max_iter=10000, random_state=42,\n",
       "                                           solver=&#x27;liblinear&#x27;),\n",
       "              n_iter=10, random_state=1,\n",
       "              search_spaces={&#x27;C&#x27;: Real(low=0.0001, high=10, prior=&#x27;log-uniform&#x27;, transform=&#x27;normalize&#x27;),\n",
       "                             &#x27;penalty&#x27;: [&#x27;l1&#x27;, &#x27;l2&#x27;]})</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-16\" type=\"checkbox\" ><label for=\"sk-estimator-id-16\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>best_estimator_: LogisticRegression</div></div></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(C=2.4458265756334576, max_iter=10000, penalty=&#x27;l1&#x27;,\n",
       "                   random_state=42, solver=&#x27;liblinear&#x27;)</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-17\" type=\"checkbox\" ><label for=\"sk-estimator-id-17\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>LogisticRegression</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.LogisticRegression.html\">?<span>Documentation for LogisticRegression</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(C=2.4458265756334576, max_iter=10000, penalty=&#x27;l1&#x27;,\n",
       "                   random_state=42, solver=&#x27;liblinear&#x27;)</pre></div> </div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "BayesSearchCV(estimator=LogisticRegression(max_iter=10000, random_state=42,\n",
       "                                           solver='liblinear'),\n",
       "              n_iter=10, random_state=1,\n",
       "              search_spaces={'C': Real(low=0.0001, high=10, prior='log-uniform', transform='normalize'),\n",
       "                             'penalty': ['l1', 'l2']})"
      ]
     },
     "execution_count": 167,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.int = int\n",
    "opt = BayesSearchCV(\n",
    "    LogisticRegression(solver=\"liblinear\", random_state=42, max_iter=10000),\n",
    "    distributions,\n",
    "    n_iter=10,\n",
    "    random_state=1\n",
    ")\n",
    "\n",
    "opt.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see the best parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('C', 2.4458265756334576), ('penalty', 'l1')])"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt.best_params_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And, as before, we can access the best model with:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-4 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-4 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-4 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-4 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-4 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-4 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-4 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-4 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-4 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-4 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-4 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-4 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-4 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-4 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-4 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-4 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-4 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-4 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-4 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-4\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>LogisticRegression(C=2.4458265756334576, max_iter=10000, penalty=&#x27;l1&#x27;,\n",
       "                   random_state=42, solver=&#x27;liblinear&#x27;)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-18\" type=\"checkbox\" checked><label for=\"sk-estimator-id-18\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>LogisticRegression</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.LogisticRegression.html\">?<span>Documentation for LogisticRegression</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(C=2.4458265756334576, max_iter=10000, penalty=&#x27;l1&#x27;,\n",
       "                   random_state=42, solver=&#x27;liblinear&#x27;)</pre></div> </div></div></div></div>"
      ],
      "text/plain": [
       "LogisticRegression(C=2.4458265756334576, max_iter=10000, penalty='l1',\n",
       "                   random_state=42, solver='liblinear')"
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9666666666666667"
      ]
     },
     "execution_count": 172,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Practice question\n",
    "Using sklearn's  make_regression, generate a toy dataset:\n",
    "\n",
    "X, y = make_regression(n_samples=100, n_features=10, noise=20,  n_informative=5,  random_state=1), \n",
    "\n",
    "split it into train and test. Then use Ridge regression with grid search cross validation to find the value of $\\alpha$ that minimizes the mean square error. Then evaluate this model on the test set. Hint: check the `scoring` parameter and the score method  *'neg_root_mean_squared_error'*."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Note"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have mentioned that no modeling decision should be done on the test dataset, and that we use validation data to make model selection and parameter tuning. Let's think about the preprocessing steps such as scaling. If we can use the `fit` method only on the train dataset, and not the validation dataset, how do we apply CV, without applying `fit` to the testing fold?\n",
    "\n",
    "Next, we will introduce `pipelines` which will help us with this issue. But without `pipelines`, in the above code, we did not apply the scaling stage before cross validation, even for the regularized models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Algorithm chains and pipelines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following exercise is adapted from Chapter 6 of *Introduction to Machine Learning with Python* by Andreas C. Müller, Sarah Guido.\n",
    "\n",
    "In this section we will review how to chain together many different processing steps and machine learning models by using *Pipeline* class. \n",
    " \n",
    "Let's start with an example of using a scaler, before applying a machine learning model, this time to do regression. We will use house price dataset from `sklearn`, and apply Ridge Regression. Before applying the model, we will split the data into train and test, and fit the `MinMax` scaler on the training data, and then transform both the training and the testing data. Note that we are scaling here to ensure that the gradient descent moves smoothly towards the minima and that the steps for gradient descent are updated at the same rate for all the features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ".. _california_housing_dataset:\n",
      "\n",
      "California Housing dataset\n",
      "--------------------------\n",
      "\n",
      "**Data Set Characteristics:**\n",
      "\n",
      ":Number of Instances: 20640\n",
      "\n",
      ":Number of Attributes: 8 numeric, predictive attributes and the target\n",
      "\n",
      ":Attribute Information:\n",
      "    - MedInc        median income in block group\n",
      "    - HouseAge      median house age in block group\n",
      "    - AveRooms      average number of rooms per household\n",
      "    - AveBedrms     average number of bedrooms per household\n",
      "    - Population    block group population\n",
      "    - AveOccup      average number of household members\n",
      "    - Latitude      block group latitude\n",
      "    - Longitude     block group longitude\n",
      "\n",
      ":Missing Attribute Values: None\n",
      "\n",
      "This dataset was obtained from the StatLib repository.\n",
      "https://www.dcc.fc.up.pt/~ltorgo/Regression/cal_housing.html\n",
      "\n",
      "The target variable is the median house value for California districts,\n",
      "expressed in hundreds of thousands of dollars ($100,000).\n",
      "\n",
      "This dataset was derived from the 1990 U.S. census, using one row per census\n",
      "block group. A block group is the smallest geographical unit for which the U.S.\n",
      "Census Bureau publishes sample data (a block group typically has a population\n",
      "of 600 to 3,000 people).\n",
      "\n",
      "A household is a group of people residing within a home. Since the average\n",
      "number of rooms and bedrooms in this dataset are provided per household, these\n",
      "columns may take surprisingly large values for block groups with few households\n",
      "and many empty houses, such as vacation resorts.\n",
      "\n",
      "It can be downloaded/loaded using the\n",
      ":func:`sklearn.datasets.fetch_california_housing` function.\n",
      "\n",
      ".. rubric:: References\n",
      "\n",
      "- Pace, R. Kelley and Ronald Barry, Sparse Spatial Autoregressions,\n",
      "  Statistics and Probability Letters, 33 (1997) 291-297\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import Ridge\n",
    "from sklearn.metrics import classification_report\n",
    "\n",
    "# load the data\n",
    "from sklearn.datasets import fetch_california_housing\n",
    "housing = fetch_california_housing()\n",
    "print(housing.DESCR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-5 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-5 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-5 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-5 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-5 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-5 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-5 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-5 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-5 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-5 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-5 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-5 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-5 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-5 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-5 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-5 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-5 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-5 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-5 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-5\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Ridge(alpha=0.01)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-19\" type=\"checkbox\" checked><label for=\"sk-estimator-id-19\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>Ridge</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.Ridge.html\">?<span>Documentation for Ridge</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>Ridge(alpha=0.01)</pre></div> </div></div></div></div>"
      ],
      "text/plain": [
       "Ridge(alpha=0.01)"
      ]
     },
     "execution_count": 179,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(housing.data, housing.target, random_state=0)\n",
    "# compute minimum and maximum on the training data\n",
    "scaler = MinMaxScaler().fit(X_train)\n",
    "# rescale the training data\n",
    "X_train_scaled = scaler.transform(X_train)\n",
    "X_test_scaled = scaler.transform(X_test)\n",
    "\n",
    "\n",
    "ridge = Ridge(alpha=0.01)\n",
    "ridge.fit(X_train_scaled, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean squared error: 0.540554646703696\n",
      "Root mean squared error: 0.7352242152593289\n",
      "R2:  0.5910622391340353\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error, root_mean_squared_error\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "y_pred=ridge.predict(X_test_scaled)\n",
    "print('Mean squared error:', mean_squared_error(y_test, y_pred))\n",
    "print('Root mean squared error:', root_mean_squared_error(y_test, y_pred))\n",
    "print('R2: ', r2_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " For `RidgeRegression` the default scoring is the R2. Hence, if we were only interested in R2, we could have simply used the *score* function:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5910622391340353"
      ]
     },
     "execution_count": 182,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ridge.score(X_test_scaled, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's say we want to find better parameters for the ridge using *GridSearchCV*. \n",
    "   \n",
    "A naive and WRONG approach to doing a grid search with data scaling might look like this:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-6 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-6 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-6 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-6 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-6 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-6 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-6 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-6 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-6 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-6 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-6 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-6 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-6 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-6 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-6 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-6 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-6 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-6 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-6 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-6\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>GridSearchCV(cv=5, estimator=Ridge(),\n",
       "             param_grid={&#x27;alpha&#x27;: [0.005, 0.01, 0.05, 0.1, 0.5, 1]},\n",
       "             scoring=&#x27;neg_root_mean_squared_error&#x27;)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-20\" type=\"checkbox\" ><label for=\"sk-estimator-id-20\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>GridSearchCV</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.model_selection.GridSearchCV.html\">?<span>Documentation for GridSearchCV</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>GridSearchCV(cv=5, estimator=Ridge(),\n",
       "             param_grid={&#x27;alpha&#x27;: [0.005, 0.01, 0.05, 0.1, 0.5, 1]},\n",
       "             scoring=&#x27;neg_root_mean_squared_error&#x27;)</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-21\" type=\"checkbox\" ><label for=\"sk-estimator-id-21\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>best_estimator_: Ridge</div></div></label><div class=\"sk-toggleable__content fitted\"><pre>Ridge(alpha=0.1)</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-22\" type=\"checkbox\" ><label for=\"sk-estimator-id-22\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>Ridge</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.Ridge.html\">?<span>Documentation for Ridge</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>Ridge(alpha=0.1)</pre></div> </div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "GridSearchCV(cv=5, estimator=Ridge(),\n",
       "             param_grid={'alpha': [0.005, 0.01, 0.05, 0.1, 0.5, 1]},\n",
       "             scoring='neg_root_mean_squared_error')"
      ]
     },
     "execution_count": 184,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "# for illustration purposes only, don't use this code!\n",
    "param_grid = {'alpha': [0.005, 0.01, 0.05, 0.1, 0.5, 1]  }\n",
    "grid = GridSearchCV(Ridge(), param_grid=param_grid, cv=5, scoring='neg_root_mean_squared_error')\n",
    "grid.fit(X_train_scaled, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, we ran the grid search over the parameter of *Ridge* using the scaled data. However, there is a subtle catch in what we just did. When scaling the data, we used ALL the data in the training set to compute the minimum and maximum of the data. We then used the scaled training data to run our grid search using cross-validation. For each split in the cross-validation, some part of the original training set will be declared the training part of the split, and some the test part of the split. The test part is used to measure the performance of a model trained on the training part when applied to new data. However, we already used the information contained in the test part of the split, when scaling the data. \n",
    "\n",
    "So, the splits in the cross-validation no longer correctly mirror how new data will look to the modeling process. We already leaked information from these parts of the data into our modeling process. This will lead to overly optimistic results during cross-validation, and possibly the selection of sub-optimal parameters.\n",
    "\n",
    "To get around this problem, the splitting of the dataset during cross-validation should be done BEFORE doing any pre-processing. Any process that extracts knowledge from the dataset should only ever be learned from the training portion of the dataset, and therefore be contained inside the cross-validation loop.\n",
    "\n",
    "To achieve this in `sklearn` with the `GridSearchCV` function, we can use the `Pipeline` class. The `Pipeline` class is a class that allows \"gluing\" together multiple processing steps into a single `sklearn` estimator.\n",
    "\n",
    "The `Pipeline` class itself has `fit`, `predict`, and `score` methods and behaves just like any other model in `sklearn`. The most common use case of the `Pipeline` class is in chaining pre-processing steps (like scaling of the data) together with a supervised model like a classifier."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Building Pipelines\n",
    " First, we build a pipeline object by providing it with a list of steps. Each step is a tuple containing a name (any string of our choosing) and an instance of an estimator:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "pipe = Pipeline([('scaler', MinMaxScaler()), \n",
    "                 ('ridge',  Ridge(alpha=0.01))])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Just like for `ColumnTransformer`, when the `sklearn` display setting is set to diagram, with `set_config(display='text')`, we can visualize the pipeline. Note that clicking on the diagram below allows us to see the details of each step:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-7 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-7 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-7 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-7 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-7 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-7 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-7 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-7 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-7 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-7 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-7 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-7 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-7 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-7 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-7 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-7 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-7 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-7 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-7 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-7\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler()), (&#x27;ridge&#x27;, Ridge(alpha=0.01))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-23\" type=\"checkbox\" ><label for=\"sk-estimator-id-23\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>Pipeline</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link \">i<span>Not fitted</span></span></div></label><div class=\"sk-toggleable__content \"><pre>Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler()), (&#x27;ridge&#x27;, Ridge(alpha=0.01))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-24\" type=\"checkbox\" ><label for=\"sk-estimator-id-24\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>MinMaxScaler</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.MinMaxScaler.html\">?<span>Documentation for MinMaxScaler</span></a></div></label><div class=\"sk-toggleable__content \"><pre>MinMaxScaler()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-25\" type=\"checkbox\" ><label for=\"sk-estimator-id-25\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>Ridge</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.Ridge.html\">?<span>Documentation for Ridge</span></a></div></label><div class=\"sk-toggleable__content \"><pre>Ridge(alpha=0.01)</pre></div> </div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('scaler', MinMaxScaler()), ('ridge', Ridge(alpha=0.01))])"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, we created two steps: the first, called \"scaler\", is an instance of *MinMaxScaler*, and the second, called \"ridge\", is an instance of *Ridge*. Now, we can fit the pipeline, like any other `sklearn` estimator:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-8 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-8 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-8 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-8 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-8 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-8 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-8 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-8 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-8 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-8 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-8 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-8 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-8 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-8 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-8 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-8 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-8 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-8 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-8 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-8\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler()), (&#x27;ridge&#x27;, Ridge(alpha=0.01))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-26\" type=\"checkbox\" ><label for=\"sk-estimator-id-26\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>Pipeline</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;scaler&#x27;, MinMaxScaler()), (&#x27;ridge&#x27;, Ridge(alpha=0.01))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-27\" type=\"checkbox\" ><label for=\"sk-estimator-id-27\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>MinMaxScaler</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.MinMaxScaler.html\">?<span>Documentation for MinMaxScaler</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>MinMaxScaler()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-28\" type=\"checkbox\" ><label for=\"sk-estimator-id-28\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>Ridge</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.Ridge.html\">?<span>Documentation for Ridge</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>Ridge(alpha=0.01)</pre></div> </div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('scaler', MinMaxScaler()), ('ridge', Ridge(alpha=0.01))])"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, *pipe.fit* first calls fit on the first step (the scaler), then transforms the training data using the scaler, and finally fits the *Ridge* with the scaled data. To evaluate on the test data with the default scoring metric of the regressor, we simply call *pipe.score*:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5910622391340353"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Calling the `score` method on the pipeline first transforms the test data using the scaler, and then calls the `score` method on the `Ridge` using the scaled test data. As we can see, the result is identical to the one we got from the code by doing the transformations by hand. Using the pipeline, we reduced the code needed for our \"preprocessing + classification\" process. The main benefit of using the pipeline, however, is that we can now use this single estimator in *GridSearchCV*."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using a pipeline in a grid search works the same way as using any other estimator. We define a parameter grid to search over, and construct a *GridSearchCV* from the pipeline and the parameter grid. When specifying the parameter grid, there is a slight change, though. We need to specify for each parameter which step of the pipeline it belongs to. The parameter that we want to adjust, *alpha*, is the parameter of *Ridge*, the second step. We gave this step the name \"ridge\". The syntax to define a parameter grid for a pipeline is to specify for each parameter the step name, followed by __ (a double underscore), followed by the parameter name. To search over the *alpha* parameter of \"ridge\" we therefore have to use \"ridge__max_alpha\" as the key in the parameter grid dictionary.\n",
    "\n",
    "We could also use the method `get_params` to see the names of all the parameters of an estimator:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'memory': None,\n",
       " 'steps': [('scaler', MinMaxScaler()), ('ridge', Ridge(alpha=0.01))],\n",
       " 'transform_input': None,\n",
       " 'verbose': False,\n",
       " 'scaler': MinMaxScaler(),\n",
       " 'ridge': Ridge(alpha=0.01),\n",
       " 'scaler__clip': False,\n",
       " 'scaler__copy': True,\n",
       " 'scaler__feature_range': (0, 1),\n",
       " 'ridge__alpha': 0.01,\n",
       " 'ridge__copy_X': True,\n",
       " 'ridge__fit_intercept': True,\n",
       " 'ridge__max_iter': None,\n",
       " 'ridge__positive': False,\n",
       " 'ridge__random_state': None,\n",
       " 'ridge__solver': 'auto',\n",
       " 'ridge__tol': 0.0001}"
      ]
     },
     "execution_count": 196,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.get_params()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see what parameters of the pipeline that we built can be tuned. Parameters of the scaling step start with *scaler__*, parameters of ridge start with *ridge__*. Note that *scaler* and *ridge* are just names we gave to the steps of the pipeline, we could have given any other name."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's define the parameters that we wish to tune:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [],
   "source": [
    "param_grid = {'ridge__alpha': [0.005, 0.01, 0.05, 0.1, 0.5, 1]}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best cross-validation score (negative RMSE):  -0.7243104320018825\n",
      "Test set score:  -0.7363007117320173\n",
      "Best parameters:  {'ridge__alpha': 0.1}\n"
     ]
    }
   ],
   "source": [
    "grid = GridSearchCV(pipe, param_grid=param_grid, cv=5, scoring='neg_root_mean_squared_error')\n",
    "grid.fit(X_train, y_train)\n",
    "y_pred=grid.predict(X_test)\n",
    "\n",
    "print('Best cross-validation score (negative RMSE): ', grid.best_score_)\n",
    "print('Test set score: ', grid.score(X_test, y_test))\n",
    "print('Best parameters: ', grid.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In contrast to the grid search we did before, now for each split in the cross-validation, the *MinMaxScaler* is refit with ONLY the training splits and no information is leaked from the test split into the parameter search."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this case since scaling is used to ensure smooth convergence of the gradient descent, there is not much difference when using the wrong and correct way of scaling. However, in many cases, if data leakage occurs it may actually lead to choosing sub-optimal parameters."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The General Pipeline Interface\n",
    "The `Pipeline` class is not restricted to pre-processing and classification or regression, but can, in fact, join any number of estimators together. For example, we could build a pipeline containing feature extraction, feature selection, scaling, and classification, for a total of four steps. Similarly, the last step could clustering instead of classification.\n",
    "\n",
    "The only requirement for estimators in a pipeline is that all but the last step need to have a `transform` method, so they can produce a new representation of the data that can be used in the next step.\n",
    "\n",
    "Internally, during the call to `Pipeline.fit`, the pipeline calls `fit` and then `transform` on each step in turn, with the input given by the output of the transform method of the previous step. For the last step in the pipeline, just `fit` is called."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pipeline is also adjusted for adequate use of sklearn's estimators and transformers. For example, when `TargetEncoder` is a part of a Pipeline and the pipeline is fitted, the pipeline will correctly call `TargetEncoder.fit_transform` and use cross fitting when encoding the training data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Often, we will want to inspect attributes of one of the steps of the pipeline, say, the coefficients of a linear model. The easiest way to access the steps in a pipeline is via the *named_steps* attribute, which is a dictionary from the step names to the estimators:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipe = Pipeline( [('scaler', StandardScaler()), ('ridge', Ridge(alpha=0.01))])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can check the steps of the pipeline:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('scaler', StandardScaler()), ('ridge', Ridge(alpha=0.01))]"
      ]
     },
     "execution_count": 208,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.steps"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualize the pipeline:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-9 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-9 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-9 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-9 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-9 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-9 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-9 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-9 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-9 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-9 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-9 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-9 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-9 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-9 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-9 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-9 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-9 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-9 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-9 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-9\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;scaler&#x27;, StandardScaler()), (&#x27;ridge&#x27;, Ridge(alpha=0.01))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-29\" type=\"checkbox\" ><label for=\"sk-estimator-id-29\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>Pipeline</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link \">i<span>Not fitted</span></span></div></label><div class=\"sk-toggleable__content \"><pre>Pipeline(steps=[(&#x27;scaler&#x27;, StandardScaler()), (&#x27;ridge&#x27;, Ridge(alpha=0.01))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-30\" type=\"checkbox\" ><label for=\"sk-estimator-id-30\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>StandardScaler</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></div></label><div class=\"sk-toggleable__content \"><pre>StandardScaler()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-31\" type=\"checkbox\" ><label for=\"sk-estimator-id-31\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>Ridge</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.Ridge.html\">?<span>Documentation for Ridge</span></a></div></label><div class=\"sk-toggleable__content \"><pre>Ridge(alpha=0.01)</pre></div> </div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('scaler', StandardScaler()), ('ridge', Ridge(alpha=0.01))])"
      ]
     },
     "execution_count": 210,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Or get pipe parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'memory': None,\n",
       " 'steps': [('scaler', StandardScaler()), ('ridge', Ridge(alpha=0.01))],\n",
       " 'transform_input': None,\n",
       " 'verbose': False,\n",
       " 'scaler': StandardScaler(),\n",
       " 'ridge': Ridge(alpha=0.01),\n",
       " 'scaler__copy': True,\n",
       " 'scaler__with_mean': True,\n",
       " 'scaler__with_std': True,\n",
       " 'ridge__alpha': 0.01,\n",
       " 'ridge__copy_X': True,\n",
       " 'ridge__fit_intercept': True,\n",
       " 'ridge__max_iter': None,\n",
       " 'ridge__positive': False,\n",
       " 'ridge__random_state': None,\n",
       " 'ridge__solver': 'auto',\n",
       " 'ridge__tol': 0.0001}"
      ]
     },
     "execution_count": 212,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.get_params()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can access a certain step of the `pipeline`, using `named_steps`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-10 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-10 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-10 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-10 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-10 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-10 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-10 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-10 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-10 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-10 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-10 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-10 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-10 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-10 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-10 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-10 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-10 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-10 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-10 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-10\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Ridge(alpha=0.01)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-32\" type=\"checkbox\" checked><label for=\"sk-estimator-id-32\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>Ridge</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.Ridge.html\">?<span>Documentation for Ridge</span></a><span class=\"sk-estimator-doc-link \">i<span>Not fitted</span></span></div></label><div class=\"sk-toggleable__content \"><pre>Ridge(alpha=0.01)</pre></div> </div></div></div></div>"
      ],
      "text/plain": [
       "Ridge(alpha=0.01)"
      ]
     },
     "execution_count": 214,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.named_steps['ridge']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we will fit the pipeline with the housing data. Since we wantto pass the feature names to the pipeline, we will convert the numpy array to a dataframe:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_pipe = pd.DataFrame(housing.data, columns=housing.feature_names)\n",
    "y_pipe = housing.target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ridge coefficiets:  [ 0.82961904  0.1187523  -0.26552558  0.30569451 -0.00450277 -0.0393263\n",
      " -0.89987946 -0.87053475]\n"
     ]
    }
   ],
   "source": [
    "# fit the pipeline defined before to the housing dataset\n",
    "pipe.fit(X_pipe, y_pipe)\n",
    "# extract the coefficients from the \"ridge\" step\n",
    "ridge_coefficients = pipe.named_steps[\"ridge\"].coef_\n",
    "print('ridge coefficiets: ', ridge_coefficients)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In another example, we could access the names of the features that are going out of the scaler step, and into the ridge step."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['MedInc', 'HouseAge', 'AveRooms', 'AveBedrms', 'Population',\n",
       "       'AveOccup', 'Latitude', 'Longitude'], dtype=object)"
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.named_steps[\"scaler\"].get_feature_names_out()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pipeline and ColumnTransformer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One of the steps in the `pipeline` can also be `ColumnTransformer`, and `pipeline` can also be a step in `ColumnTransformer`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's go back to the census dataset, but this time will import the csv with the dataset that has missing values:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>capital-gain-category</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>occupation</th>\n",
       "      <th>income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39.0</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>NaN</td>\n",
       "      <td>40.0</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50.0</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>cat1</td>\n",
       "      <td>13.0</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38.0</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>cat1</td>\n",
       "      <td>40.0</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53.0</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>cat1</td>\n",
       "      <td>40.0</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28.0</td>\n",
       "      <td>Private</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>cat1</td>\n",
       "      <td>40.0</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    age         workclass  education capital-gain-category  hours-per-week  \\\n",
       "0  39.0         State-gov  Bachelors                   NaN            40.0   \n",
       "1  50.0  Self-emp-not-inc  Bachelors                  cat1            13.0   \n",
       "2  38.0           Private    HS-grad                  cat1            40.0   \n",
       "3  53.0           Private       11th                  cat1            40.0   \n",
       "4  28.0           Private  Bachelors                  cat1            40.0   \n",
       "\n",
       "          occupation income  \n",
       "0       Adm-clerical  <=50K  \n",
       "1    Exec-managerial  <=50K  \n",
       "2  Handlers-cleaners  <=50K  \n",
       "3  Handlers-cleaners  <=50K  \n",
       "4     Prof-specialty  <=50K  "
      ]
     },
     "execution_count": 223,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('census_data_missing.csv' )\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age                      1628\n",
       "workclass                1628\n",
       "education                1628\n",
       "capital-gain-category    6189\n",
       "hours-per-week           1628\n",
       "occupation               1628\n",
       "income                      0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 224,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Recall that we used one hot encoding for feature 'workclass', target encoding for 'occupation', ordinal encoding for 'education' and 'capital-gain-category'. Now let us first impute the missing values of categorical features with mode, and of continuous with the median. Then, we will encode the categorical features, scale all with the StandardScaler and then train the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-11 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-11 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-11 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-11 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-11 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-11 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-11 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-11 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-11 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-11 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-11 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-11 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-11 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-11 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-11 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-11 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-11 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-11 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-11 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-11\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;preprocessor&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;numerical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_median&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;median&#x27;))]),\n",
       "                                                  [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                                 (&#x27;categorical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                                  (&#x27;onehot&#x27;,\n",
       "                                                                   OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                                                                 handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                                                                 min_frequ...\n",
       "                                                                   OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                                                               &#x27;1st-4th&#x27;,\n",
       "                                                                                               &#x27;5th-6th&#x27;,\n",
       "                                                                                               &#x27;7th-8th&#x27;,\n",
       "                                                                                               &#x27;9th&#x27;,\n",
       "                                                                                               &#x27;10th&#x27;,\n",
       "                                                                                               &#x27;11th&#x27;,\n",
       "                                                                                               &#x27;12th&#x27;,\n",
       "                                                                                               &#x27;HS-grad&#x27;,\n",
       "                                                                                               &#x27;Prof-school&#x27;,\n",
       "                                                                                               &#x27;Assoc-acdm&#x27;,\n",
       "                                                                                               &#x27;Assoc-voc&#x27;,\n",
       "                                                                                               &#x27;Some-college&#x27;,\n",
       "                                                                                               &#x27;Bachelors&#x27;,\n",
       "                                                                                               &#x27;Masters&#x27;,\n",
       "                                                                                               &#x27;Doctorate&#x27;],\n",
       "                                                                                              [&#x27;cat1&#x27;,\n",
       "                                                                                               &#x27;cat2&#x27;,\n",
       "                                                                                               &#x27;cat3&#x27;,\n",
       "                                                                                               &#x27;cat4&#x27;]]))]),\n",
       "                                                  [&#x27;education&#x27;,\n",
       "                                                   &#x27;capital-gain-category&#x27;])])),\n",
       "                (&#x27;scaler&#x27;, StandardScaler()),\n",
       "                (&#x27;classifier&#x27;, LogisticRegression(solver=&#x27;liblinear&#x27;))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-33\" type=\"checkbox\" ><label for=\"sk-estimator-id-33\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>Pipeline</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link \">i<span>Not fitted</span></span></div></label><div class=\"sk-toggleable__content \"><pre>Pipeline(steps=[(&#x27;preprocessor&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;numerical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_median&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;median&#x27;))]),\n",
       "                                                  [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                                 (&#x27;categorical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                                  (&#x27;onehot&#x27;,\n",
       "                                                                   OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                                                                 handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                                                                 min_frequ...\n",
       "                                                                   OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                                                               &#x27;1st-4th&#x27;,\n",
       "                                                                                               &#x27;5th-6th&#x27;,\n",
       "                                                                                               &#x27;7th-8th&#x27;,\n",
       "                                                                                               &#x27;9th&#x27;,\n",
       "                                                                                               &#x27;10th&#x27;,\n",
       "                                                                                               &#x27;11th&#x27;,\n",
       "                                                                                               &#x27;12th&#x27;,\n",
       "                                                                                               &#x27;HS-grad&#x27;,\n",
       "                                                                                               &#x27;Prof-school&#x27;,\n",
       "                                                                                               &#x27;Assoc-acdm&#x27;,\n",
       "                                                                                               &#x27;Assoc-voc&#x27;,\n",
       "                                                                                               &#x27;Some-college&#x27;,\n",
       "                                                                                               &#x27;Bachelors&#x27;,\n",
       "                                                                                               &#x27;Masters&#x27;,\n",
       "                                                                                               &#x27;Doctorate&#x27;],\n",
       "                                                                                              [&#x27;cat1&#x27;,\n",
       "                                                                                               &#x27;cat2&#x27;,\n",
       "                                                                                               &#x27;cat3&#x27;,\n",
       "                                                                                               &#x27;cat4&#x27;]]))]),\n",
       "                                                  [&#x27;education&#x27;,\n",
       "                                                   &#x27;capital-gain-category&#x27;])])),\n",
       "                (&#x27;scaler&#x27;, StandardScaler()),\n",
       "                (&#x27;classifier&#x27;, LogisticRegression(solver=&#x27;liblinear&#x27;))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-34\" type=\"checkbox\" ><label for=\"sk-estimator-id-34\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>preprocessor: ColumnTransformer</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for preprocessor: ColumnTransformer</span></a></div></label><div class=\"sk-toggleable__content \"><pre>ColumnTransformer(transformers=[(&#x27;numerical&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;imputation_median&#x27;,\n",
       "                                                  SimpleImputer(strategy=&#x27;median&#x27;))]),\n",
       "                                 [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                (&#x27;categorical&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                  SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                 (&#x27;onehot&#x27;,\n",
       "                                                  OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                                                handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                                                min_frequency=0.01,\n",
       "                                                                sparse_output=False))]...\n",
       "                                 Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                  SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                 (&#x27;onehot&#x27;,\n",
       "                                                  OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                                              &#x27;1st-4th&#x27;,\n",
       "                                                                              &#x27;5th-6th&#x27;,\n",
       "                                                                              &#x27;7th-8th&#x27;,\n",
       "                                                                              &#x27;9th&#x27;,\n",
       "                                                                              &#x27;10th&#x27;,\n",
       "                                                                              &#x27;11th&#x27;,\n",
       "                                                                              &#x27;12th&#x27;,\n",
       "                                                                              &#x27;HS-grad&#x27;,\n",
       "                                                                              &#x27;Prof-school&#x27;,\n",
       "                                                                              &#x27;Assoc-acdm&#x27;,\n",
       "                                                                              &#x27;Assoc-voc&#x27;,\n",
       "                                                                              &#x27;Some-college&#x27;,\n",
       "                                                                              &#x27;Bachelors&#x27;,\n",
       "                                                                              &#x27;Masters&#x27;,\n",
       "                                                                              &#x27;Doctorate&#x27;],\n",
       "                                                                             [&#x27;cat1&#x27;,\n",
       "                                                                              &#x27;cat2&#x27;,\n",
       "                                                                              &#x27;cat3&#x27;,\n",
       "                                                                              &#x27;cat4&#x27;]]))]),\n",
       "                                 [&#x27;education&#x27;, &#x27;capital-gain-category&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-35\" type=\"checkbox\" ><label for=\"sk-estimator-id-35\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>numerical</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;age&#x27;, &#x27;hours-per-week&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-36\" type=\"checkbox\" ><label for=\"sk-estimator-id-36\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content \"><pre>SimpleImputer(strategy=&#x27;median&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-37\" type=\"checkbox\" ><label for=\"sk-estimator-id-37\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>categorical</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;workclass&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-38\" type=\"checkbox\" ><label for=\"sk-estimator-id-38\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content \"><pre>SimpleImputer(strategy=&#x27;most_frequent&#x27;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-39\" type=\"checkbox\" ><label for=\"sk-estimator-id-39\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>OneHotEncoder</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></div></label><div class=\"sk-toggleable__content \"><pre>OneHotEncoder(drop=&#x27;first&#x27;, handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "              min_frequency=0.01, sparse_output=False)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-40\" type=\"checkbox\" ><label for=\"sk-estimator-id-40\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>cat_target</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;occupation&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-41\" type=\"checkbox\" ><label for=\"sk-estimator-id-41\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content \"><pre>SimpleImputer(strategy=&#x27;most_frequent&#x27;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-42\" type=\"checkbox\" ><label for=\"sk-estimator-id-42\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>TargetEncoder</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.TargetEncoder.html\">?<span>Documentation for TargetEncoder</span></a></div></label><div class=\"sk-toggleable__content \"><pre>TargetEncoder(target_type=&#x27;binary&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-43\" type=\"checkbox\" ><label for=\"sk-estimator-id-43\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>ordinal</div></div></label><div class=\"sk-toggleable__content \"><pre>[&#x27;education&#x27;, &#x27;capital-gain-category&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-44\" type=\"checkbox\" ><label for=\"sk-estimator-id-44\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content \"><pre>SimpleImputer(strategy=&#x27;most_frequent&#x27;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-45\" type=\"checkbox\" ><label for=\"sk-estimator-id-45\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>OrdinalEncoder</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.OrdinalEncoder.html\">?<span>Documentation for OrdinalEncoder</span></a></div></label><div class=\"sk-toggleable__content \"><pre>OrdinalEncoder(categories=[[&#x27;Preschool&#x27;, &#x27;1st-4th&#x27;, &#x27;5th-6th&#x27;, &#x27;7th-8th&#x27;, &#x27;9th&#x27;,\n",
       "                            &#x27;10th&#x27;, &#x27;11th&#x27;, &#x27;12th&#x27;, &#x27;HS-grad&#x27;, &#x27;Prof-school&#x27;,\n",
       "                            &#x27;Assoc-acdm&#x27;, &#x27;Assoc-voc&#x27;, &#x27;Some-college&#x27;,\n",
       "                            &#x27;Bachelors&#x27;, &#x27;Masters&#x27;, &#x27;Doctorate&#x27;],\n",
       "                           [&#x27;cat1&#x27;, &#x27;cat2&#x27;, &#x27;cat3&#x27;, &#x27;cat4&#x27;]])</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-46\" type=\"checkbox\" ><label for=\"sk-estimator-id-46\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>StandardScaler</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></div></label><div class=\"sk-toggleable__content \"><pre>StandardScaler()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator  sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-47\" type=\"checkbox\" ><label for=\"sk-estimator-id-47\" class=\"sk-toggleable__label  sk-toggleable__label-arrow\"><div><div>LogisticRegression</div></div><div><a class=\"sk-estimator-doc-link \" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.LogisticRegression.html\">?<span>Documentation for LogisticRegression</span></a></div></label><div class=\"sk-toggleable__content \"><pre>LogisticRegression(solver=&#x27;liblinear&#x27;)</pre></div> </div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('preprocessor',\n",
       "                 ColumnTransformer(transformers=[('numerical',\n",
       "                                                  Pipeline(steps=[('imputation_median',\n",
       "                                                                   SimpleImputer(strategy='median'))]),\n",
       "                                                  ['age', 'hours-per-week']),\n",
       "                                                 ('categorical',\n",
       "                                                  Pipeline(steps=[('imputation_mode',\n",
       "                                                                   SimpleImputer(strategy='most_frequent')),\n",
       "                                                                  ('onehot',\n",
       "                                                                   OneHotEncoder(drop='first',\n",
       "                                                                                 handle_unknown='infrequent_if_exist',\n",
       "                                                                                 min_frequ...\n",
       "                                                                   OrdinalEncoder(categories=[['Preschool',\n",
       "                                                                                               '1st-4th',\n",
       "                                                                                               '5th-6th',\n",
       "                                                                                               '7th-8th',\n",
       "                                                                                               '9th',\n",
       "                                                                                               '10th',\n",
       "                                                                                               '11th',\n",
       "                                                                                               '12th',\n",
       "                                                                                               'HS-grad',\n",
       "                                                                                               'Prof-school',\n",
       "                                                                                               'Assoc-acdm',\n",
       "                                                                                               'Assoc-voc',\n",
       "                                                                                               'Some-college',\n",
       "                                                                                               'Bachelors',\n",
       "                                                                                               'Masters',\n",
       "                                                                                               'Doctorate'],\n",
       "                                                                                              ['cat1',\n",
       "                                                                                               'cat2',\n",
       "                                                                                               'cat3',\n",
       "                                                                                               'cat4']]))]),\n",
       "                                                  ['education',\n",
       "                                                   'capital-gain-category'])])),\n",
       "                ('scaler', StandardScaler()),\n",
       "                ('classifier', LogisticRegression(solver='liblinear'))])"
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.impute import SimpleImputer\n",
    "numeric_preprocessor = Pipeline([ (\"imputation_median\", SimpleImputer( strategy=\"median\")) ])\n",
    "\n",
    "categorical_preprocessor = Pipeline([\n",
    "        (\"imputation_mode\", SimpleImputer( strategy=\"most_frequent\")),\n",
    "        (\"onehot\", OneHotEncoder(sparse_output=False, drop='first', \n",
    "                              min_frequency=0.01, handle_unknown='infrequent_if_exist'))\n",
    "    ])\n",
    "\n",
    "ordinal_preprocessor = Pipeline([\n",
    "        (\"imputation_mode\", SimpleImputer( strategy=\"most_frequent\")),\n",
    "        (\"onehot\", OrdinalEncoder(categories=categories))\n",
    "    ])\n",
    "\n",
    "target_enc_preprocessor = Pipeline([\n",
    "        (\"imputation_mode\", SimpleImputer( strategy=\"most_frequent\")),\n",
    "        ('target_enc', TargetEncoder(target_type='binary')) \n",
    "    ])\n",
    "        \n",
    "preprocessor = ColumnTransformer([\n",
    "        (\"numerical\", numeric_preprocessor, ['age', 'hours-per-week']),\n",
    "        (\"categorical\", categorical_preprocessor, ['workclass']),\n",
    "        (\"cat_target\", target_enc_preprocessor, ['occupation']),\n",
    "        (\"ordinal\", ordinal_preprocessor, ['education','capital-gain-category']),\n",
    "    ])\n",
    "\n",
    "pipe = Pipeline([\n",
    "     ('preprocessor', preprocessor),\n",
    "     ('scaler', StandardScaler()),  \n",
    "     ('classifier',  LogisticRegression(solver='liblinear'))])\n",
    "\n",
    "pipe"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that, while `ColumnTransformer` applies the assigned transformations simultaneously, the `Pipeline` executes its steps sequentially."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(df.drop(columns='income'), df['income'], random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-12 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-12 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-12 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-12 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-12 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-12 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-12 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-12 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-12 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-12 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-12 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-12 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-12 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-12 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-12 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-12 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-12 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-12 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-12 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-12\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;preprocessor&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;numerical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_median&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;median&#x27;))]),\n",
       "                                                  [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                                 (&#x27;categorical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                                  (&#x27;onehot&#x27;,\n",
       "                                                                   OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                                                                 handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                                                                 min_frequ...\n",
       "                                                                   OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                                                               &#x27;1st-4th&#x27;,\n",
       "                                                                                               &#x27;5th-6th&#x27;,\n",
       "                                                                                               &#x27;7th-8th&#x27;,\n",
       "                                                                                               &#x27;9th&#x27;,\n",
       "                                                                                               &#x27;10th&#x27;,\n",
       "                                                                                               &#x27;11th&#x27;,\n",
       "                                                                                               &#x27;12th&#x27;,\n",
       "                                                                                               &#x27;HS-grad&#x27;,\n",
       "                                                                                               &#x27;Prof-school&#x27;,\n",
       "                                                                                               &#x27;Assoc-acdm&#x27;,\n",
       "                                                                                               &#x27;Assoc-voc&#x27;,\n",
       "                                                                                               &#x27;Some-college&#x27;,\n",
       "                                                                                               &#x27;Bachelors&#x27;,\n",
       "                                                                                               &#x27;Masters&#x27;,\n",
       "                                                                                               &#x27;Doctorate&#x27;],\n",
       "                                                                                              [&#x27;cat1&#x27;,\n",
       "                                                                                               &#x27;cat2&#x27;,\n",
       "                                                                                               &#x27;cat3&#x27;,\n",
       "                                                                                               &#x27;cat4&#x27;]]))]),\n",
       "                                                  [&#x27;education&#x27;,\n",
       "                                                   &#x27;capital-gain-category&#x27;])])),\n",
       "                (&#x27;scaler&#x27;, StandardScaler()),\n",
       "                (&#x27;classifier&#x27;, LogisticRegression(solver=&#x27;liblinear&#x27;))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-48\" type=\"checkbox\" ><label for=\"sk-estimator-id-48\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>Pipeline</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;preprocessor&#x27;,\n",
       "                 ColumnTransformer(transformers=[(&#x27;numerical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_median&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;median&#x27;))]),\n",
       "                                                  [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                                 (&#x27;categorical&#x27;,\n",
       "                                                  Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                                   SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                                  (&#x27;onehot&#x27;,\n",
       "                                                                   OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                                                                 handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                                                                 min_frequ...\n",
       "                                                                   OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                                                               &#x27;1st-4th&#x27;,\n",
       "                                                                                               &#x27;5th-6th&#x27;,\n",
       "                                                                                               &#x27;7th-8th&#x27;,\n",
       "                                                                                               &#x27;9th&#x27;,\n",
       "                                                                                               &#x27;10th&#x27;,\n",
       "                                                                                               &#x27;11th&#x27;,\n",
       "                                                                                               &#x27;12th&#x27;,\n",
       "                                                                                               &#x27;HS-grad&#x27;,\n",
       "                                                                                               &#x27;Prof-school&#x27;,\n",
       "                                                                                               &#x27;Assoc-acdm&#x27;,\n",
       "                                                                                               &#x27;Assoc-voc&#x27;,\n",
       "                                                                                               &#x27;Some-college&#x27;,\n",
       "                                                                                               &#x27;Bachelors&#x27;,\n",
       "                                                                                               &#x27;Masters&#x27;,\n",
       "                                                                                               &#x27;Doctorate&#x27;],\n",
       "                                                                                              [&#x27;cat1&#x27;,\n",
       "                                                                                               &#x27;cat2&#x27;,\n",
       "                                                                                               &#x27;cat3&#x27;,\n",
       "                                                                                               &#x27;cat4&#x27;]]))]),\n",
       "                                                  [&#x27;education&#x27;,\n",
       "                                                   &#x27;capital-gain-category&#x27;])])),\n",
       "                (&#x27;scaler&#x27;, StandardScaler()),\n",
       "                (&#x27;classifier&#x27;, LogisticRegression(solver=&#x27;liblinear&#x27;))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-49\" type=\"checkbox\" ><label for=\"sk-estimator-id-49\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>preprocessor: ColumnTransformer</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for preprocessor: ColumnTransformer</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(transformers=[(&#x27;numerical&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;imputation_median&#x27;,\n",
       "                                                  SimpleImputer(strategy=&#x27;median&#x27;))]),\n",
       "                                 [&#x27;age&#x27;, &#x27;hours-per-week&#x27;]),\n",
       "                                (&#x27;categorical&#x27;,\n",
       "                                 Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                  SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                 (&#x27;onehot&#x27;,\n",
       "                                                  OneHotEncoder(drop=&#x27;first&#x27;,\n",
       "                                                                handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "                                                                min_frequency=0.01,\n",
       "                                                                sparse_output=False))]...\n",
       "                                 Pipeline(steps=[(&#x27;imputation_mode&#x27;,\n",
       "                                                  SimpleImputer(strategy=&#x27;most_frequent&#x27;)),\n",
       "                                                 (&#x27;onehot&#x27;,\n",
       "                                                  OrdinalEncoder(categories=[[&#x27;Preschool&#x27;,\n",
       "                                                                              &#x27;1st-4th&#x27;,\n",
       "                                                                              &#x27;5th-6th&#x27;,\n",
       "                                                                              &#x27;7th-8th&#x27;,\n",
       "                                                                              &#x27;9th&#x27;,\n",
       "                                                                              &#x27;10th&#x27;,\n",
       "                                                                              &#x27;11th&#x27;,\n",
       "                                                                              &#x27;12th&#x27;,\n",
       "                                                                              &#x27;HS-grad&#x27;,\n",
       "                                                                              &#x27;Prof-school&#x27;,\n",
       "                                                                              &#x27;Assoc-acdm&#x27;,\n",
       "                                                                              &#x27;Assoc-voc&#x27;,\n",
       "                                                                              &#x27;Some-college&#x27;,\n",
       "                                                                              &#x27;Bachelors&#x27;,\n",
       "                                                                              &#x27;Masters&#x27;,\n",
       "                                                                              &#x27;Doctorate&#x27;],\n",
       "                                                                             [&#x27;cat1&#x27;,\n",
       "                                                                              &#x27;cat2&#x27;,\n",
       "                                                                              &#x27;cat3&#x27;,\n",
       "                                                                              &#x27;cat4&#x27;]]))]),\n",
       "                                 [&#x27;education&#x27;, &#x27;capital-gain-category&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-50\" type=\"checkbox\" ><label for=\"sk-estimator-id-50\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>numerical</div></div></label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;age&#x27;, &#x27;hours-per-week&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-51\" type=\"checkbox\" ><label for=\"sk-estimator-id-51\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>SimpleImputer(strategy=&#x27;median&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-52\" type=\"checkbox\" ><label for=\"sk-estimator-id-52\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>categorical</div></div></label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;workclass&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-53\" type=\"checkbox\" ><label for=\"sk-estimator-id-53\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>SimpleImputer(strategy=&#x27;most_frequent&#x27;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-54\" type=\"checkbox\" ><label for=\"sk-estimator-id-54\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>OneHotEncoder</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(drop=&#x27;first&#x27;, handle_unknown=&#x27;infrequent_if_exist&#x27;,\n",
       "              min_frequency=0.01, sparse_output=False)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-55\" type=\"checkbox\" ><label for=\"sk-estimator-id-55\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>cat_target</div></div></label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;occupation&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-56\" type=\"checkbox\" ><label for=\"sk-estimator-id-56\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>SimpleImputer(strategy=&#x27;most_frequent&#x27;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-57\" type=\"checkbox\" ><label for=\"sk-estimator-id-57\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>TargetEncoder</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.TargetEncoder.html\">?<span>Documentation for TargetEncoder</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>TargetEncoder(target_type=&#x27;binary&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-58\" type=\"checkbox\" ><label for=\"sk-estimator-id-58\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>ordinal</div></div></label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;education&#x27;, &#x27;capital-gain-category&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-59\" type=\"checkbox\" ><label for=\"sk-estimator-id-59\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>SimpleImputer</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.impute.SimpleImputer.html\">?<span>Documentation for SimpleImputer</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>SimpleImputer(strategy=&#x27;most_frequent&#x27;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-60\" type=\"checkbox\" ><label for=\"sk-estimator-id-60\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>OrdinalEncoder</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.OrdinalEncoder.html\">?<span>Documentation for OrdinalEncoder</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>OrdinalEncoder(categories=[[&#x27;Preschool&#x27;, &#x27;1st-4th&#x27;, &#x27;5th-6th&#x27;, &#x27;7th-8th&#x27;, &#x27;9th&#x27;,\n",
       "                            &#x27;10th&#x27;, &#x27;11th&#x27;, &#x27;12th&#x27;, &#x27;HS-grad&#x27;, &#x27;Prof-school&#x27;,\n",
       "                            &#x27;Assoc-acdm&#x27;, &#x27;Assoc-voc&#x27;, &#x27;Some-college&#x27;,\n",
       "                            &#x27;Bachelors&#x27;, &#x27;Masters&#x27;, &#x27;Doctorate&#x27;],\n",
       "                           [&#x27;cat1&#x27;, &#x27;cat2&#x27;, &#x27;cat3&#x27;, &#x27;cat4&#x27;]])</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-61\" type=\"checkbox\" ><label for=\"sk-estimator-id-61\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>StandardScaler</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-62\" type=\"checkbox\" ><label for=\"sk-estimator-id-62\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>LogisticRegression</div></div><div><a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.6/modules/generated/sklearn.linear_model.LogisticRegression.html\">?<span>Documentation for LogisticRegression</span></a></div></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(solver=&#x27;liblinear&#x27;)</pre></div> </div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('preprocessor',\n",
       "                 ColumnTransformer(transformers=[('numerical',\n",
       "                                                  Pipeline(steps=[('imputation_median',\n",
       "                                                                   SimpleImputer(strategy='median'))]),\n",
       "                                                  ['age', 'hours-per-week']),\n",
       "                                                 ('categorical',\n",
       "                                                  Pipeline(steps=[('imputation_mode',\n",
       "                                                                   SimpleImputer(strategy='most_frequent')),\n",
       "                                                                  ('onehot',\n",
       "                                                                   OneHotEncoder(drop='first',\n",
       "                                                                                 handle_unknown='infrequent_if_exist',\n",
       "                                                                                 min_frequ...\n",
       "                                                                   OrdinalEncoder(categories=[['Preschool',\n",
       "                                                                                               '1st-4th',\n",
       "                                                                                               '5th-6th',\n",
       "                                                                                               '7th-8th',\n",
       "                                                                                               '9th',\n",
       "                                                                                               '10th',\n",
       "                                                                                               '11th',\n",
       "                                                                                               '12th',\n",
       "                                                                                               'HS-grad',\n",
       "                                                                                               'Prof-school',\n",
       "                                                                                               'Assoc-acdm',\n",
       "                                                                                               'Assoc-voc',\n",
       "                                                                                               'Some-college',\n",
       "                                                                                               'Bachelors',\n",
       "                                                                                               'Masters',\n",
       "                                                                                               'Doctorate'],\n",
       "                                                                                              ['cat1',\n",
       "                                                                                               'cat2',\n",
       "                                                                                               'cat3',\n",
       "                                                                                               'cat4']]))]),\n",
       "                                                  ['education',\n",
       "                                                   'capital-gain-category'])])),\n",
       "                ('scaler', StandardScaler()),\n",
       "                ('classifier', LogisticRegression(solver='liblinear'))])"
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7940056504114974"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.score(X_test,y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We should also do a grid search and fine tune our model: it is also possible to search over the actual steps being performed in the pipeline (say whether to use *StandardScaler* or *MinMaxScaler*). We could also test different imputing strategies, and model parameters. We can also grid search different classifiers, as we will see in the next section. But note, trying all possible solutions is usually not a viable machine learning strategy."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Grid-Searching Which Model To Use\n",
    "\n",
    "Here is an example comparing a *Ridge*, *Lasso* and plain Linear Regression on the housing dataset.\n",
    "\n",
    "We start by defining the pipeline. Here, we explicitly name the steps. We want two steps, one for the pre-processing and then a regressor. We can instantiate this using  *Ridge* and *MinMaxScaler*:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipe = Pipeline([('preprocessing', MinMaxScaler()), \n",
    "                 ('regressor', Ridge())])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we can define the parameter_grid to search over. We want the regressor to be either *Ridge*, *Lasso* or *LinearRegression*. Because they have different parameters to tune, and may benefit from different pre-processing, we can make use of the list of search grids. To assign an estimator to a step, we use the name of the step as the parameter name. When we want to skip a step in the pipeline (for example, because we may not need pre-processing for the *LinearRegression*), we can set that step to *None*. Note that *GridSearchCV* allows the param_grid to be a list of dictionaries. Each dictionary in the list is expanded into an independent grid. If we are using *RandomizedSearchCV* with a list of dictionaries, first a dictionary is sampled uniformly, and then a parameter is sampled using that dictionary."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's see the pipe parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'memory': None,\n",
       " 'steps': [('preprocessing', MinMaxScaler()), ('regressor', Ridge())],\n",
       " 'transform_input': None,\n",
       " 'verbose': False,\n",
       " 'preprocessing': MinMaxScaler(),\n",
       " 'regressor': Ridge(),\n",
       " 'preprocessing__clip': False,\n",
       " 'preprocessing__copy': True,\n",
       " 'preprocessing__feature_range': (0, 1),\n",
       " 'regressor__alpha': 1.0,\n",
       " 'regressor__copy_X': True,\n",
       " 'regressor__fit_intercept': True,\n",
       " 'regressor__max_iter': None,\n",
       " 'regressor__positive': False,\n",
       " 'regressor__random_state': None,\n",
       " 'regressor__solver': 'auto',\n",
       " 'regressor__tol': 0.0001}"
      ]
     },
     "execution_count": 236,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe.get_params()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import Lasso,  LinearRegression\n",
    "\n",
    "param_grid = [\n",
    "    {'regressor': [Ridge()], 'preprocessing': [MinMaxScaler(),StandardScaler()],\n",
    "     'regressor__alpha': [0.01, 0.03, 0.05, 0.1, 0.3,  0.5, 1 ] },\n",
    "    {'regressor': [Lasso()], 'preprocessing': [ MinMaxScaler(),StandardScaler()], \n",
    "     'regressor__alpha': [0.01, 0.03, 0.05, 0.1, 0.3,  0.5, 1  ]},\n",
    " {'regressor': [LinearRegression()], 'preprocessing': [None]}\n",
    "]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we can instantiate and run the grid search on the housing dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best params:  {'preprocessing': MinMaxScaler(), 'regressor': Ridge(), 'regressor__alpha': 0.1}\n",
      "Best cross-validation score  -0.7243104320018825\n",
      "Test-set score:   -0.7363007117320173\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(housing.data, housing.target, random_state=0)\n",
    "\n",
    "grid = GridSearchCV(pipe, param_grid, cv=5, scoring='neg_root_mean_squared_error', n_jobs=-1)\n",
    "grid.fit(X_train, y_train)\n",
    "\n",
    "print('Best params: ', grid.best_params_)\n",
    "print('Best cross-validation score ', grid.best_score_)\n",
    "print('Test-set score:  ', grid.score(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The outcome of the grid search is that *Ridge* with *MinMaxScaler* preprocessing, *alpha=0.1*  gave the best result."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Splitting data in cross-validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "During Cross-Validation, the data is split into k-folds. To specify how the data will be split, so far we have only passed the number of cross validation folds in the parameter `cv`. The default behavior of `GridSearchCV` and `RandomizedSearchCV`, when we pass the number of folds when the estimator is a classifier and y is either binary or multiclass, is the use of stratified stratified k-fold cross-validation. This will enforce the class distribution in each split of the data to match the distribution in the complete training dataset. If the data was just split without considering stratification, in the case of severe class imbalance, one or more folds could end up with a few or no examples from the minority class. This means that some or perhaps many of the model evaluations would be misleading, as the model need only predict the majority class correctly."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hence, by default `StratifiedKFold`cross-validator is used that provides train/holdout indices to split data. The folds are made by preserving the percentage of samples for each class, without shuffling the data. Let's illustrate the workings of this cross-validator using the small imbalanced dataset we generated previously."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import make_classification\n",
    "\n",
    "X, y = make_classification(n_samples=1000, n_features=2, n_informative=2,\n",
    "                            n_redundant=0, n_repeated=0, n_classes=3,\n",
    "                            n_clusters_per_class=1,\n",
    "                            weights=[0.01, 0.05, 0.94],\n",
    "                            class_sep=0.8, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2    0.932\n",
       "1    0.055\n",
       "0    0.013\n",
       "Name: proportion, dtype: float64"
      ]
     },
     "execution_count": 245,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(y).value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`Split` method of this cross-validator generates indices to split data into training and holdout set. Let's see if we wanted to do 5-fold cross validation, what would be the class distribution for each holdout fold."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fold 0: Class distribution of validation instances\n",
      "2    0.935\n",
      "1    0.055\n",
      "0    0.010\n",
      "Name: proportion, dtype: float64\n",
      "Fold 1: Class distribution of validation instances\n",
      "2    0.935\n",
      "1    0.055\n",
      "0    0.010\n",
      "Name: proportion, dtype: float64\n",
      "Fold 2: Class distribution of validation instances\n",
      "2    0.930\n",
      "1    0.055\n",
      "0    0.015\n",
      "Name: proportion, dtype: float64\n",
      "Fold 3: Class distribution of validation instances\n",
      "2    0.930\n",
      "1    0.055\n",
      "0    0.015\n",
      "Name: proportion, dtype: float64\n",
      "Fold 4: Class distribution of validation instances\n",
      "2    0.930\n",
      "1    0.055\n",
      "0    0.015\n",
      "Name: proportion, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import  StratifiedKFold \n",
    "skf = StratifiedKFold(n_splits=5)\n",
    "for fold, (train_idx, val_idx) in enumerate(skf.split(X, y)):\n",
    "    print(f'Fold {fold}: Class distribution of validation instances')\n",
    "    print(pd.Series(y[val_idx]).value_counts(normalize=True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's compare this with K-Folds cross-validator that just split dataset into k consecutive folds (without shuffling by default)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fold 0\n",
      "Class distribution of validation instances\n",
      "2    0.945\n",
      "1    0.050\n",
      "0    0.005\n",
      "Name: proportion, dtype: float64\n",
      "Fold 1\n",
      "Class distribution of validation instances\n",
      "2    0.935\n",
      "1    0.055\n",
      "0    0.010\n",
      "Name: proportion, dtype: float64\n",
      "Fold 2\n",
      "Class distribution of validation instances\n",
      "2    0.930\n",
      "1    0.055\n",
      "0    0.015\n",
      "Name: proportion, dtype: float64\n",
      "Fold 3\n",
      "Class distribution of validation instances\n",
      "2    0.92\n",
      "1    0.07\n",
      "0    0.01\n",
      "Name: proportion, dtype: float64\n",
      "Fold 4\n",
      "Class distribution of validation instances\n",
      "2    0.930\n",
      "1    0.045\n",
      "0    0.025\n",
      "Name: proportion, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import KFold \n",
    "kf = KFold(n_splits=5)\n",
    "for  fold, (train_idx, val_idx) in enumerate(kf.split(X, y)):\n",
    "    print(f'Fold {fold}')\n",
    "    print('Class distribution of validation instances')\n",
    "    print(pd.Series(y[val_idx]).value_counts(normalize=True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, the distribution of the rarest class varies in the validation folds."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Another important detail which we have not discussed so far is the shuffling of the data instances when the folds are made. By default, data is not shuffled, hence running cross-validation twice with the default parameters, will give exactly the same results. However, if the ordering of data instances in the dataset is not arbitrary, for example data points with the same class label are contiguous, shuffling it first may be essential to get a meaningful cross-validation result. In this case we can use `ShuffleSplit` and `StratifiedShuffleSplit`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Another common case of data that is ordered is time series. Usually today's stock price is correlated with yesterday's and tomorrow's. In applications, we usually have data up to some point, and then try to make predictions for the future, in other words, we are using past data to predict the future data. For this case we can use `TimeSeriesSplit` which simulates that, by taking increasing chunks of data from the past and making predictions on the next chunk.  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now to visualize the workings of all this splits, we will use the code from: https://scikit-learn.org/stable/auto_examples/model_selection/plot_cv_indices.html\n",
    "\n",
    "Here, to goal is only to analyze the obtained plots with the code (no need to analyze or replicate the code)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import   ShuffleSplit, StratifiedShuffleSplit,   TimeSeriesSplit\n",
    "from matplotlib.patches import Patch\n",
    "cmap_data = plt.cm.Paired\n",
    "cmap_cv = plt.cm.coolwarm\n",
    "n_splits = 5\n",
    "cvs = [\n",
    "    KFold,\n",
    "    ShuffleSplit,\n",
    "    StratifiedKFold,\n",
    "    StratifiedShuffleSplit,\n",
    "    TimeSeriesSplit\n",
    "]\n",
    "X = np.random.randn(100, 10)\n",
    "percentiles_classes = [0.1, 0.3, 0.6]\n",
    "y = np.hstack([[ii] * int(100 * perc) for ii, perc in enumerate(percentiles_classes)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_cv_indices(cv, X, y, ax, n_splits, lw=10):\n",
    "    \"\"\"Create a sample plot for indices of a cross-validation object.\"\"\"\n",
    "\n",
    "    # Generate the training/testing visualizations for each CV split\n",
    "    for ii, (tr, tt) in enumerate(cv.split(X=X, y=y )):\n",
    "        # Fill in indices with the training/test groups\n",
    "        indices = np.array([np.nan] * len(X))\n",
    "        indices[tt] = 1\n",
    "        indices[tr] = 0\n",
    "\n",
    "        # Visualize the results\n",
    "        ax.scatter(\n",
    "            range(len(indices)),\n",
    "            [ii + 0.5] * len(indices),\n",
    "            c=indices,\n",
    "            marker=\"_\",\n",
    "            lw=lw,\n",
    "            cmap=cmap_cv,\n",
    "            vmin=-0.2,\n",
    "            vmax=1.2,\n",
    "        )\n",
    "\n",
    "    # Plot the data classes and groups at the end\n",
    "    ax.scatter(\n",
    "        range(len(X)), [ii + 1.5] * len(X), c=y, marker=\"_\", lw=lw, cmap=cmap_data\n",
    "    )\n",
    "\n",
    "\n",
    "    # Formatting\n",
    "    yticklabels = list(range(n_splits)) + [\"class\"]\n",
    "    ax.set(\n",
    "        yticks=np.arange(n_splits + 1) + 0.5,\n",
    "        yticklabels=yticklabels,\n",
    "        xlabel=\"Data sample index\",\n",
    "        ylabel=\"CV Iteration\",\n",
    "        ylim=[n_splits + 2.2, -0.2],\n",
    "        xlim=[0, 100],\n",
    "    )\n",
    "    ax.set_title(\"{}\".format(type(cv).__name__), fontsize=15)\n",
    "    return ax"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAj4AAAEiCAYAAAD5ztjRAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjAsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvlHJYcgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAM3BJREFUeJzt3QmczeUex/HfMGbGOvZ9SchOUrqoKApJlkSuGym6lsl2FSpbroYUia4uFSV7kXIjspNdtow125WtbFlmM//7+j2cc8+ZzZmMOcv/8369TmfOOf8585x/45zvPM/veZ4gy7IsAQAAsIFM3m4AAABARiH4AAAA2yD4AAAA2yD4AAAA2yD4AAAA2yD4AAAA2yD4AAAA2yD4AAAA2yD4AAAA2yD4ABkgKCjIXJITHx8vbdu2NY+XLVtWjh8/bu6fOnWq8/tSuugxf5bj+YcOHZrm13LXXXf96Z8LAN4U7NWfDtichp527drJl19+KeXKlZMVK1ZIsWLF3I4pU6aMPPTQQ8l+vwYlAIDnCD6Al8TFxclzzz0n8+bNk/Lly8vy5culaNGiSY7T0HM7PTsAgP8j+ABeCj1t2rSRr7/+WipWrGhCT+HChb3dLAAIeNT4ABksNjZWnn32WRN6KleuLCtXrky30PPdd9/J448/Lnny5JGwsDDTkzRgwAC5cOFCmp7n3LlzEhERYXqg9HkqVaok48aNE8uy0qWdAOAt9PgAGRx6WrduLd9++61Uq1ZNfvjhBylQoEC6PHdkZKS8/vrrEhwcLPXq1ZP8+fPLunXrZNSoUTJ//nxZvXq1FCpU6JbPc/78eTO8FhUVZQJZ8+bNTRDq16+fHDx4MF3aCgDeQvABMlCrVq3kP//5j9x7770m9OTLly9dnnfz5s3y5ptvSo4cOczzPvjgg+b+mJgYef7552Xu3LnSo0cPU0R9KxqeNPQ0btxYvvrqK8mWLZu5f9OmTdKgQYN0aS8AeAtDXUAG0tCj08GnTZvmcej57LPPkp3KXr9+fecxEyZMkISEBHnllVecoUeFhoaax7JmzWp6fRxT5VNy5coV8/MyZcpkvs8RelStWrVMeAIAf0aPD5CB6tata4afdN2eVatWmeGoW0lpOnuFChWcX69Zs8Zct2/fPslxBQsWlCeeeEIWLFhgfrbOJEvJ1q1b5dq1aybk6M9NTKfe69AZAPgrgg+QwT0+jz32mGzbts0MJelsrly5cqX6PZ5MZ//111/NdUoLCzruP3HihEfPU6pUqVSfBwD8FUNdQAYKDw+X77//3syS0t6Vpk2bytWrV+/4z01p1WgAsBuCD5DBdHhLC5B1KGnt2rXSsmVLM9vrdjgWPjx69Giyjx85csRcJ14VOrEiRYqk+jwp3Q8A/oLgA3iBBgwNP8WLF5clS5aYupvr16//6ed7+OGHzfXMmTOTPHb27FnTy6S9PlpjlJqaNWuaQmjtjfrll1+SPD5r1qw/3UYA8AUEH8BLtF5Gw48WH+uMqxdeeOFPLxCos610JtYHH3wgW7Zscd6vPUk600sLlnUqfYkSJVJ9Hp0Or9PfNYQ5vs9Bn1dnegGAPyP4AF6kKytrj4+utPzFF1/86eniOgtr+PDhcunSJaldu7ZZvVlnYOkmprNnzzYboH744YceL4So7dJVoHU4TnujGjVqZJ63Q4cOf6p9AOArCD6Al1WvXl0WLVpkelsmTpwor7322p96Hl14cOHChWbVZl3QUDc/1XV89Pk2btzo0arNKm/evGbae7du3UwPlG6tcezYMRk5cqSMHz/+T7UNAHxFkMXmOwAAwCbo8QEAALbBAoYAAJ+kRfZxcXHebgb8QJYsWSRz5sweHUvwAQD4FK3AOHXqlFy4cMHbTYEfyZ07txQuXPiWC7YSfAAAPsURenSpB90ol5XHcaugrCvgnzlzxm0h1pQQfAAAPjW85Qg9+fLl83Zz4Cd04VWl4Ud/d1Ib9qK4GQDgMxw1PdrTA6SF43fmVnVh9Ph4ICEhwexanTNnTrpcASAdhib++OMPs8ecrjieHN5rkVae/s4QfDygoedWS/0DANLm+PHjZr86JN3Opnfv3uaS2oe8bnXTokUL8UYN1vPPPy8//vijmU3lSRH60KFDzWKo27dvT/EY3bZHn0uPu5MIPh7Qnh7HP9JcuXJ5uzkA4Nd0axX9Y9Lx3uqJhIvnJOHqZckombLlkEzheT0+PqUP7ZUrV8qjjz4q58+fN7OO/D1wqbFjx8rJkydNiAkPDxd/Q/BJQ/eZhh6CDwBk7NCEhp6LE4eIXI+XDJM5WMK7DUtT+LGLQ4cOSc2aNc0egP6I4mYAgE8zPT0ZGXrU9fg71sP01VdfSeXKlc1eetrL8t5776V6/IEDB+SRRx6RsLAwqVSpkixdujTJMbt27ZLHHnvMzG7S2XAvv/yyXL78//bXr18/SU9OixYtTE+V4/GjR49Knz59TCBNKZRqe7X9n3/+uTnG8f26n1/z5s3NnoPaQdCmTRs5ffp0qrP3+vbta3rBtL26p2BG7aBF8AEAIINs3brVhILnnnvOhBWtfRk0aJBMnTo1xck1rVq1kpCQELPZ8EcffST9+/d3O+bKlSvSqFEjyZMnj9mgeO7cufLDDz9IRESEx+2aN2+eqbd66623zDCWXpKjz9+4cWPzGvSYcePGmTZq6Dl37pysWrXKBLNffvlF2rZtm+LP07Cnr/nTTz+VtWvXmu/VmqWMwFAXAADpYOHChabHI3HPhqsxY8ZIgwYNTNhR99xzj+zZs0dGjx7t7D1xpQFm79698v3335tZcOrtt9+WJk2aOI+ZMWOGREdHm16Y7Nmzm/smTJggzZo1k1GjRkmhQoVu2fa8efOatW+07kpXP05JgQIFTE+V9iw5jtOgoyHu8OHDzolA2hbt1dKg9MADDyR5nvfff18GDhxoQp3SQKevMSPQ4wMAQDrQImYt+HW9fPzxx27HREVFSd26dd3u09s6nJU4JDmO1zDhCD2qdu3aSY6pXr26M/Q4nlN7Yvbt2yd3WtTNNrrOftYhOR3G0scSu3jxouktevDBB533BQcHy/3333/H22p+Vob8FAAAApwGj7Jly7rd99///ld8ga6XlLiGJs6mG8DS4wMAQAapWLGirFu3zu0+va1DXslts6DH61IqrjU3GzZsSHLMjh07TK2P63Nq2ClfvrxziMr1Oa5fvy67d+92ex6tI0qu18mT16Rt1IuDDt/p9H7t+UlMp8Drflpas+QQHx9v6p8yAsEHAIAM8o9//EOWLVsmw4cPl/3798tnn31m6nH69euX7PENGzY0oahjx44m3KxZs0beeOMNt2Pat29vZnzpMRpmVqxYIa+88opZZNBR36Mzvv7zn/+Yi9YMdevWLcnCgzpja/Xq1XLixAn57bffPH5N2saqVauadmzbtk02bdokHTp0kHr16qU4fNWrVy8ZOXKkWfdI29O9e3ePFkJMDwQfAAAyyH333Sdz5syRWbNmSZUqVWTw4MFmJlVyhc1Ke210ttO1a9ekVq1a0rlzZxkxYkSSPaq0MFhnRmkhcevWrU0BtQYqhxdffNEEow43A8ndd99tapJcaTuOHDkiZcqUMT1EntJp7QsWLDCzynTavQYhff7Zs2enGgA1mGmbtGZJi6pbtmwpGSHIyqiJ836+yqh2zWlBFgsYAsCde0/V2Uk6O6h06dKmF0OxgCE8kdzvTnIobk6D//b9m+QMySJ2NLrEeOfXsUE3f6HYRBBAGoSF3LiOib6Upu/T8KEhxJe3rID/IPjAI7GZsnq7CQD8XHTsjevYm9dpoSGEIIL0QI0PAACwDdsEnw8//NBUrOu4ny6apFXnAADAXmwRfLSyXDdDGzJkiJlqpytc6r4mZ86c8XbTAABABrJF8NG9Ubp06SKdOnUyiynpniA6/U83RwMAAPYR8MEnNjbWrAap6wq4rougt9evX5/s98TExJjplq4XAADg/wI++Ojqk7oEd+LdafX2qVOnkv2eyMhIs8aE4+K68RoAAPBfAR98/oyBAweahbUcF9f9RwAAgP8K+OCTP39+s/Hb6dOn3e7X24ULF072e0JDQ81qoq4XAAAyks5Efv/99z0+fuXKlWb7iIzY8+rrr782O9Hr52vv3r3T7fVo+/W576SAX8BQd5utWbOm2RSuRYsW5r6EhARzOyIiwtvNAwB44PRvMXLxUsZtWRGeK1gK5Q/16Fj9sE6NzigeOnRomtuwefNmyZ49u8fH16lTx+zAriUad9rf//53M2GoZ8+eZp8tfxLwwUfpVHbdCE13idVN3jRxXrlyxfxPAwD4fujp0GuHxMZl3NaSIVmC5PNx1T0KPxo2XJdP0Y1H9+3b57wvR44czq91e0ytOw0OvvXHb1o2CjVtDglJcSQjPV2+fNksB6PLwhQtWlT8TcAPdam2bdvKu+++a34Z7733Xtm+fbssXrw4ScEzAMD3aE9PRoYepT/P0x4mDRuOi/a2aA+Q4/bevXtNj8iiRYvM6IOWUqxdu1YOHTokzZs3N59DGox0V/Uffvgh1aEhfd6PP/7Y7GKuS7KUK1dOvvnmmxSHuqZOnSq5c+c2O7dXrFjR/JzGjRu7BbX4+HjTa6PH5cuXT/r37286ChwjJInpz3D08Dz22GPm5+l96quvvpLKlSub16htf++991I9bwcOHDC7uevCwrrUzNKlS93/H8TGmpGZIkWKmGNKlSplJh/dLlsEH6Un7+jRo2aq+saNG83qzfBcSMI150Us68YFANK4SaleQm9uVmonAwYMkJEjR0pUVJRUq1bN9Jo8+eSTpuzip59+MoGkWbNmcuzYsVSfZ9iwYdKmTRvZuXOn+f727dvLuXPnUjz+6tWr5g//adOmyerVq83z9+vXz/n4qFGjZPr06TJlyhRZt26dWb4ltRobHU5z9GZp0NEQpffpsjHarueee0527dplhvYGDRpkwldytOSkVatWppdKP5N1fT0NXa4++OADE+zmzJljfqa2UwPV7bLFUFd6KT7mC9sWOn/g7QYACBiXLmWWj14XW3nrrbfk8ccfd97Omzev2UXAYfjw4TJ//nzzQZ9a/ekLL7wg7dq1M1+//fbbJhzoFkwanJITFxdnQkWZMmXM7YiICNMWh/Hjx5uZzNqLpCZMmCDfffddij9fg0rBggWdr8ExtKYLBTdo0MCEHXXPPffInj17ZPTo0abNiWnvlvaGaW+UY7hMX0+TJk2cx2hI016thx56yPQsaY9PerBNjw8AAN6iNaautMdHe150CEqHmXQYSnuDbtXjo71FDlr4rH+Mp7b9kg6JOUKPKlKkiPN4Xa5FZzhr7auDztLSIbm00rbXrVvX7T69rcNZWtOU3PG6Rp5rjVDt2rXdjtHApKUp5cuXN8NxS5YskfRA8AEA4A5LPDtLQ4/28Ggvx5o1a8wHfNWqVU1dS2qyZMnidlt7QnTYKC3HW35SqnDffffJ4cOHTW/YtWvXzFBa69atb/t5CT4AAGQwrafRHg0dYtLAo0NGR44cydA2hIeHm+JqnTbvoL0zupl3WmnPlb4mV3pbh7y0Fym543VxYNdC6w0bNiQ5Tnu0dILS5MmTzYw5rStKrabJE9T4AACQwbR2Zd68eaagWXthtDYmtZ6bO+WVV14xM6V0McIKFSqYmp/z58/fcm2ixP7xj3+YmWnaO6NBRffC1Hqhf/3rX8ker/tlaijSGWRaB6RF1W+88YbbMVo3pENzNWrUMHtszp071wREHRq8HfT4AACQwfRDPU+ePGZGlIYfXRNHh3YyWv/+/U2xdIcOHUyNjdYaaVt0+nhaaNt19tWsWbOkSpUqZvkYLaJOrrBZaZDRoT4dwtIao86dO8uIESPcjtFp8++8846pj9JQpT1iWnit33s7gix/GezzIk2i2iWohWB2ndUFABnxnhodHW3qOkqXLu388PX1BQwDSUJCghmG0noa7b3xJ8n97iSHoS4AgE/T8KEhxFe3rPBnR48eNbOl6tWrZ9a50+EpDQ9//etfJVARfAAAPk9DiB2CSEbLlCmTWWRQZ5npAJAOU+kaO9rrE6gIPgAA2FSJEiWSzMYKdBQ3AwAA2yD4AAAA2yD4AAB8DhOOcad+Zwg+AACf4dhiQXcVB9LC8TuTeJuOxChuToPeY89LSFjSzdbsoN+xCJGbC3mGJESb67St6xlYRpcY7/w6NujmehFpXOkUsJuwkBvXMdGXUjxGtzfQlXkdG2nqJptpXUUY9uvpuXr1qvmd0d+d5LbIcEXwgUdCrWgRep6dYjNl9XYTAL8TfXP/zVvsw2m2JVCp7ToOJKahx/G7I3YPPqtXrzZ7gWzdutVsiKbLZLdo0cLbzQIAJEN7eHSPpoIFC0pcXJy3mwM/oMNbt+rpsVXwuXLlilSvXl1efPFFadWqlbebAwDwgH6QefphBnjKFsGnSZMm5gIAAOyNWV0AAMA2bNHjk1a6UZteXHcSBgAA/o8en2RERkZKeHi486J7mQAAAP9H8EnGwIED5eLFi87L8ePHvd0kAACQDhjqSkZoaKi5AACAwGKL4HP58mU5ePCg8/bhw4dl+/btkjdvXilZsqRX2wYAADKOLYLPli1b5NFHH3Xe7tu3r7nu2LGjTJ061YstAwAAGckWwad+/frs9AsAAOwRfHD7YnQjTjYpdQpJuOb8mk1KgbRtUhqU4O2WwM6CLLpCbknX8dFp7TrDK1euXN5uDgD4Nd5T4U1MZwcAALZB8AEAALZB8AEAALZB8AEAALZB8AEAALZB8AEAALZB8AEAALbxpxYwPHDggKxYsULOnDkjCQnuK1ENHjw4vdoGAADg3eAzefJk6datm+TPn18KFy4sQS6r1erXBB8AABAwweef//ynjBgxQvr3739nWgQAAOArNT7nz5+XZ5999s60BgAAwJeCj4aeJUuW3JnWAAAA+NJQV9myZWXQoEGyYcMGqVq1qmTJksXt8Z49e0qg6j32vISEXfd2M+AD+h2LuPFFELvVq9ElxrvdNjvWs1s9UtidPSb6krebAhtLc/CZNGmS5MiRQ1atWmUurrS4OZCDD+AQat0IO2J5uyW+ITZTVm83AX4gOvbGdezNa8Avgs/hw4fvTEsAAAB8eQFDy7LMxZdFRkbKAw88IDlz5pSCBQtKixYtZN++fd5uFgAA8Jfg8/nnn5v6nqxZs5pLtWrVZNq0aeKLdDiuR48epiZp6dKlEhcXJ0888YRcuXLF200DAAC+PtQ1ZswYU9wcEREhdevWNfetXbtWunbtKr/99pv06dNHfMnixYvdbk+dOtX0/GzdulUeeeQRr7ULAAD4QfAZP368TJw4UTp06OC87+mnn5bKlSvL0KFDfS74JHbx4kVznTdv3hSPiYmJMReHS5eYgQAAgC2Huk6ePCl16tRJcr/ep4/5Mt1XrHfv3qanqkqVKqnWBYWHhzsvJUqUyNB2AgAAHwk+uo7PnDlzktw/e/ZsKVeunPgyrfXZvXu3zJo1K9XjBg4caHqGHJfjx49nWBsBAIAPDXUNGzZM2rZtK6tXr3bW+Kxbt06WLVuWbCDyFVqTtHDhQtPu4sWLp3psaGiouQAAAJv3+DzzzDOyceNGszv7119/bS769aZNm6Rly5bia3S6vYae+fPny/Lly6V06dLebhIAAPCXHh9Vs2ZN+eKLL8Qf6PDWjBkzZMGCBWYtn1OnTpn7tXZHp+IDAAD78Cj46KymXLlyOb9OjeM4X6Ez0FT9+vXd7p8yZYq88MILXmoVAADw2eCTJ08eM2NL17/JnTu32ZMruSElvf/6dd/axNPXV5aGf4rRTTgVm5QaIQnX3G6zSSlS26Q0KMHbLYGdBVkeJANd/VgLmYODg5NsTJpYvXr1JNBoL5cOjekML1/r0QIAf8N7Kny+x8c1zGhxsK5rk7jXR/MT074BAEBAzerS4HP27Nkk9587d44ZUwAAILCCj6OWJ7HLly9LWNjNugcAAAB/ns7et29fc62hRzcpzZYtm/MxLWjWtX3uvffeO9NKAACAjAw+P/30k7PHZ9euXRISEvL/GR0hIVK9enXp169ferQJAADAu8FnxYoV5rpTp04ybtw4KvEBAEDgr9ysC/8BAADYZsuKLVu2mA1Jjx07JrGxsW6PzZs3L73aBgAA4N1ZXbNmzZI6depIVFSU2fgzLi5Ofv75Z7MBqC5IBQAAEDDB5+2335axY8fKt99+a4qatd5n79690qZNGylZsuSdaSUAAIA3gs+hQ4ekadOm5msNPleuXDFT3Pv06SOTJk1KjzYBAAD4RvDRDUv/+OMP83WxYsVk9+7d5usLFy7I1atX07+FAAAA3ipufuSRR2Tp0qVStWpVefbZZ6VXr16mvkfva9CgQXq1CwAAwPvBZ8KECRIdHW2+fuONNyRLlizy448/yjPPPCNvvvmmBLLeY89LSNh1bzcD8Dn9jkXc+OLmbjYhCdGOL21rdInxzq9jg8J02Xuxu7Cb697GRF/ydlNgY2kKPvHx8bJw4UJp1KiRuZ0pUyYZMGDAnWobAD8Rat34Y0gsb7fEd8RmyurtJvic6JurnyRaBQXw3Rqf4OBg6dq1q7PHxx9MnDhRqlWrZlaa1kvt2rVl0aJF3m4WAADwh+LmWrVqyfbt28VfFC9eXEaOHClbt241Cy8+9thj0rx5c7P2EAAAsJc01/h0797d7NR+/PhxqVmzpmTPnt3tce1d8SXNmjVzuz1ixAjTC7RhwwapXLmy19oFAAD8IPg899xz5rpnz57O+3QdH921Xa+vX/fd4l9t29y5c83aQzrkBQAA7CXNwefw4cPib3bt2mWCjtYm5ciRw2y1UalSpRSPj4mJMReHS5eYgQAAgC2DT6lSpcTflC9f3tQlXbx4Ub788kvp2LGjrFq1KsXwExkZKcOGDcvwdgIAAB8rblbTpk2TunXrStGiReXo0aPmvvfff18WLFggvki31ihbtqypSdJQU716dbPHWEoGDhxoQpLjovVMAADAhsFHC4O1uPnJJ58021Q4anpy585two8/SEhIcBvKSiw0NNQ5/d1xAQAANgw+48ePl8mTJ5tVmzNnzuy8//777ze1NL5Ge29Wr14tR44cMe3T2ytXrpT27dt7u2kAAMAfiptr1KiRbC+JzpbyNWfOnJEOHTrIyZMnJTw83Ey3//777+Xxxx/3dtMAAICvB5/SpUubQuHERc6LFy+WihUriq/55JNPvN0EAADgr8FH63t69Ohhpobr2j2bNm2SmTNnmqLhjz/++M60EoBPi9FNOBWblDqFJFxzfs0mpe6blAYleLslsLMgS9NLGk2fPl2GDh0qhw4dMrd1dpdO/37ppZckEOk6PjpMpjO8KHQGgNvDeyr8Lvg4XL16VS5fviwFCxaUQMY/UgBIP7ynwq9mdekmnzqNXWXLls0ZevQXWR8DAAAImOCjU8FjY2OT3K81P2vWrEmvdgEAAHivuHnnzp3Or/fs2SOnTp1y3tZFDHVWV7FixdK/hQAAABkdfO69916z+7pekhvSypo1q1ncEAAAwO+Djy5cqHXQd999t5nCXqBAAbe9sLTWx3UlZwAAAL8NPo4FC3WfKwAAgIAOPt98841Hxz399NO30x4AAADvB58WLVrc8hit/3Hs1g4AAOC3wYchLgAAYLt1fAAAAPwVwQcAANhGmndnt7PeY89LSBg1TABurd+xCHarTyQoLKu5/iMmzttNgY0RfADgDgi1okX+9BbQgcmKvnbjOpbgAz8Y6gqU2VojR440s8969+7t7aYAAABfDT66D9eAAQNk//794q82b94s//73v6VatWrebgoAAPDl4NOjRw/58ssvpWLFivLwww/L1KlT5erVq+IvLl++LO3bt5fJkydLnjx5vN0cAADgy8Fn0KBBcvDgQVm2bJnZrysiIkKKFCkiXbp0kY0bN4qv0+DWtGlTadiwobebAgAA/GU6e/369eWzzz6TU6dOyXvvvSdRUVFSu3ZtqVy5sowZM0Z80axZs2Tbtm0SGRnp0fExMTFy6dIltwsAALDxOj45cuSQzp07y9q1a+Xbb781QejVV18VX3P8+HHp1auXTJ8+XcLCwjz6Hg1I4eHhzkuJEiXueDsBAIAPBx+t79E6n3r16pmNSfPlyycjRowQX7N161Y5c+aM3HfffRIcHGwuq1atkg8++MB8ndxstYEDB8rFixedFw1PAADAhuv4/Pjjj/Lpp5/K3LlzJT4+Xlq3bi3Dhw+XRx55RHxRgwYNZNeuXW73derUSSpUqCD9+/eXzJkzJ/me0NBQcwEAADYNPu+8845MmTLFTGe///77ZfTo0dKuXTvJmTOn+DJtX5UqVdzuy549u+mhSnw/AAAIbB4HHw06f/vb30xPD4EBAAAEdPD59ddfJUuWLBIIVq5c6e0mAAAAXy5uXrNmjVSqVCnZqd1aAKzT2fUYAIBITFCYxGS6cWHLrv9vUmouoTc2KwW8IciyLI/+TerMrUcffVT69OmT7OM6S2rFihUyf/58CTQa9nRauwa8XLlyebs5AODXeE+FX/T47NixQxo3bpzi40888YSZOg4AAOD3wef06dOp1vjomjhnz55Nr3YBAAB4d3f23bt3p/j4zp07zd5dAAAAfh98nnzySbNRaXR0dJLHrl27JkOGDJGnnnoqvdsHAACQ8cXNOtSl2z7oSse6M3v58uXN/Xv37pUPP/zQbP2gG4EWKlRIAg2FeACQfnhPhV+s46OBRrer6Natm9nLypGXgoKCpFGjRib8BGLoAQAANt2rq1SpUvLdd9/J+fPn5eDBgyb8lCtXTvLkyXPnWggAAOCtTUqVBp0HHnggvdoAAADgW8XNAAAA/o7gAwAAbIPgAwAAbIPgAwAAbONPFTfb1dwdJyRbjqS7099JrS5+4Pw6WGLNdVCGtgDJWTzlh5tfBUl8XLyXWwP4h+Cs2c311djr3m4KbIzg4+Oy3Aw78C3xcbxxA2kVf+3KjWv+/cCLfHKo68iRI2ZhxO3bt3u7KQAAIID4ZPABAAC4Ewg+AADANrwafBISEuSdd96RsmXLSmhoqJQsWVJGjBiR5DjdAPWll16S0qVLS9asWc0GqePGjXM7ZuXKlVKrVi3Jnj275M6dW+rWrStHjx41j+3YsUMeffRRyZkzp9kQr2bNmrJly5YMe50AAMA3eLW4WTc7nTx5sowdO1YeeughOXnypNntPbmAVLx4cZk7d67ky5fPbJb68ssvS5EiRaRNmzYSHx8vLVq0kC5dusjMmTMlNjZWNm3aZOqEVPv27aVGjRoyceJEs7u81g5lyZIlxXbFxMSYi+tOwgAAwP95Lfj88ccfptdmwoQJ0rFjR3NfmTJlTADS4mZXGlKGDRvmvK09P+vXr5c5c+aY4KPB5OLFi/LUU0+Z51AVK1Z0Hn/s2DF59dVXpUKFCua2bqyamsjISLefBwAAAoPXhrqioqJMr0qDBg08Ov7DDz80Q1QFChSQHDlyyKRJk0ygUXnz5pUXXnhBGjVqJM2aNTOBSnuPHPr27SudO3eWhg0bysiRI+XQoUO37InSIOW4HD9+/DZfLQAAsHXw0VodT82aNUv69etn6nyWLFlihqo6depkhrQcpkyZYnqB6tSpI7Nnz5Z77rlHNmzYYB4bOnSo/Pzzz9K0aVNZvny5VKpUSebPn5/iz9N6I60Fcr0AAAD/57Xgo8NNGn6WLVt2y2PXrVtnAk337t1NrY4WQyfXa6OPaW+N1gBVqVJFZsyY4XxMg1CfPn1McGrVqpUJSgAAwF68FnzCwsKkf//+8tprr8nnn39ugoz20HzyySfJhiSdhfX999/L/v37ZdCgQbJ582bn44cPHzaBR3t8dCaXhpsDBw6YOp9r165JRESEmfWlj2mI0u91rQECAAD24NVZXRpggoODZfDgwfLrr7+aWVpdu3ZNctzf//53+emnn6Rt27Zmpla7du1M78+iRYvM49myZTOzwT777DP5/fffzfP06NHDfJ/O+NL7OnToIKdPn5b8+fObHh+KlwEAsJ8gy7IsbzfC1+mssfDwcPl49R7JliNnhv5sNin1TWxSCtzeJqXPfrHdTB6hhhIZjeCThuDDP1IAuH28p8Kb2LICAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYBsEHAADYRrC3G+APLMsy15cuXfJ2UwDA7zneSx3vrUBGIvh44PfffzfXJUqU8HZTACCg3lvDw8O93QzYDMHHA3nz5jXXx44d4x/pzb/WNAQeP35ccuXKJXbH+XDH+UiKc+Lu4sWLUrJkSed7K5CRCD4eyJTpRimUhh7etP5PzwXn4/84H+44H0lxTpJ/bwUyEr91AADANgg+AADANgg+HggNDZUhQ4aYa3A+EuN8uON8JMU5ccf5gDcFWcwnBAAANkGPDwAAsA2CDwAAsA2CDwAAsA2Czy18+OGHctddd0lYWJg8+OCDsmnTJrGDyMhIeeCBByRnzpxSsGBBadGihezbt8/tmOjoaOnRo4fky5dPcuTIIc8884ycPn1a7GDkyJESFBQkvXv3tvX5OHHihPztb38zrzlr1qxStWpV2bJli/NxLSEcPHiwFClSxDzesGFDOXDggASi69evy6BBg6R06dLmtZYpU0aGDx/uti1DIJ+P1atXS7NmzaRo0aLm38bXX3/t9rgnr/3cuXPSvn17s9ZR7ty55aWXXpLLly9n8CtBoCP4pGL27NnSt29fM/tg27ZtUr16dWnUqJGcOXNGAt2qVavMh/iGDRtk6dKlEhcXJ0888YRcuXLFeUyfPn3k22+/lblz55rjf/31V2nVqpUEus2bN8u///1vqVatmtv9djsf58+fl7p160qWLFlk0aJFsmfPHnnvvfckT548zmPeeecd+eCDD+Sjjz6SjRs3Svbs2c2/IQ2JgWbUqFEyceJEmTBhgkRFRZnb+vrHjx9vi/Oh7w36Hql/LCbHk9euoefnn3827zkLFy40Yerll1/OwFcBW9BZXUherVq1rB49ejhvX79+3SpatKgVGRlp2c2ZM2f0z1Zr1apV5vaFCxesLFmyWHPnznUeExUVZY5Zv369Faj++OMPq1y5ctbSpUutevXqWb169bLt+ejfv7/10EMPpfh4QkKCVbhwYWv06NHO+/Q8hYaGWjNnzrQCTdOmTa0XX3zR7b5WrVpZ7du3t9350N/7+fPnO2978tr37Nljvm/z5s3OYxYtWmQFBQVZJ06cyOBXgEBGj08KYmNjZevWraY71nV5db29fv16sePeOsqxt46eG+0Fcj0/FSpUMPvvBPL50V6wpk2bur1uu56Pb775Ru6//3559tlnzXBojRo1ZPLkyc7HDx8+LKdOnXI7J7rtiw4ZB+I5qVOnjixbtkz2799vbu/YsUPWrl0rTZo0seX5cOXJa9drHd7S3ykHPV7fd7WHCEgv7NWVgt9++82M2RcqVMjtfr29d+9esZOEhARTy6LDGlWqVDH36ZtYSEiIeaNKfH70sUA0a9YsM+SpQ12J2fF8/PLLL2ZoR4eDX3/9dXNeevbsac5Dx44dna87uX9DgXhOBgwYYDYj1cCbOXNm8/4xYsQIM3yj7HY+XHny2vVaA7Sr4OBg88dWoJ8fZCyCDzzq5di9e7f569WudFftXr16mdoDLXTHjUCsf52//fbb5rb2+OjvidZwaPCxmzlz5sj06dNlxowZUrlyZdm+fbv5g0GLfe14PgBfxVBXCvLnz2/+aks8K0dvFy5cWOwiIiLCFBmuWLFCihcv7rxfz4EOB164cMEW50eHsrSo/b777jN/hepFC5i1WFO/1r9c7XQ+lM7OqVSpktt9FStWlGPHjpmvHa/bLv+GXn31VdPr89xzz5nZbc8//7wpeNcZknY8H648ee16nXjiSHx8vJnpFejnBxmL4JMC7a6vWbOmGbN3/QtXb9euXVsCndYnauiZP3++LF++3EzRdaXnRmfzuJ4fne6uH3qBeH4aNGggu3btMn/FOy7a26HDGI6v7XQ+lA59Jl7iQOtbSpUqZb7W3xn9wHI9JzoUpPUagXhOrl69aupRXOkfT/q+Ycfz4cqT167X+oeD/pHhoO89ev60FghIN96urvZls2bNMrMOpk6damYcvPzyy1bu3LmtU6dOWYGuW7duVnh4uLVy5Urr5MmTzsvVq1edx3Tt2tUqWbKktXz5cmvLli1W7dq1zcUuXGd12fF8bNq0yQoODrZGjBhhHThwwJo+fbqVLVs264svvnAeM3LkSPNvZsGCBdbOnTut5s2bW6VLl7auXbtmBZqOHTtaxYoVsxYuXGgdPnzYmjdvnpU/f37rtddes8X50BmPP/30k7noR8uYMWPM10ePHvX4tTdu3NiqUaOGtXHjRmvt2rVmBmW7du28+KoQiAg+tzB+/HjzYRYSEmKmt2/YsMGyA33jSu4yZcoU5zH6htW9e3crT5485gOvZcuWJhzZNfjY8Xx8++23VpUqVcwfCBUqVLAmTZrk9rhOYx40aJBVqFAhc0yDBg2sffv2WYHo0qVL5vdB3y/CwsKsu+++23rjjTesmJgYW5yPFStWJPueoYHQ09f++++/m6CTI0cOK1euXFanTp1MoALSE7uzAwAA26DGBwAA2AbBBwAA2AbBBwAA2AbBBwAA2AbBBwAA2AbBBwAA2AbBBwAA2AbBBwAA2AbBB7Cx+vXrmx3Eb0dQUJB8/fXXt/UcU6dOldy5c9/WcwCAJwg+CCgvvPCC+SDWi24aqrumP/744/Lpp586N4v0FB/Gnjl58qQ0adLE280AAI8QfBBwGjdubD6Mjxw5IosWLZJHH31UevXqJU899ZTEx8d7u3kBR3fdDg0N9XYzAMAjBB8EHP0Q1g/jYsWKyX333Sevv/66LFiwwIQg7cVxGDNmjFStWlWyZ88uJUqUkO7du8vly5fNYytXrpROnTrJxYsXnT1IQ4cONY9NmzZN7r//fsmZM6f5OX/961/lzJkzqbbpX//6l5QrV07CwsJML1Tr1q2djy1evFgeeugh07uUL18+E9AOHTrkfFwDnP78OXPmyMMPPyxZs2aVBx54QPbv3y+bN282bcmRI4fpdTl79qxb71eLFi1k2LBhUqBAAcmVK5d07dpVYmNjU2xnTEyM9OvXz5w7PS8PPvigOReeDnU52jpv3jwTOLNlyybVq1eX9evXu32P/n8oWbKkebxly5by+++/J3le/X+m///0nN19993mdTiC61tvvSVFixZ1+76mTZuan5nWnj0ANpOuW54CXqY7QTdv3jzZx6pXr241adLEeXvs2LHW8uXLrcOHD1vLli2zypcvb3Xr1s08pjtqv//++2aHaN1hXS+OXaI/+eQT67vvvrMOHTpkrV+/3qpdu7bb8ya2efNmK3PmzNaMGTOsI0eOWNu2bbPGjRvnfPzLL7+0vvrqK+vAgQPWTz/9ZDVr1syqWrWqdf36dfO4tk//qeru54sXL7b27Nlj/eUvf7Fq1qxp1a9f31q7dq15zrJly1pdu3Z1Oxe6y3Xbtm2t3bt3WwsXLrQKFChgvf766ynuMN+5c2erTp061urVq62DBw9ao0ePNjtp79+/P8XXp22bP39+krbqz9Pdt1u3bm2VKlXKiouLM8ds2LDBypQpkzVq1CjzuJ6L3LlzW+Hh4c7n1J+v537q1KnmPC9ZssS66667rKFDh5rH4+PjzXlv0aKFuT1hwgTzHEePHk2xnQCgCD6wTfDRAFCxYsUUv3fu3LlWvnz5nLenTJni9mGcWrDRD3tHMEpMQ41+iF+6dMmj13D27FnzfLt27XILEx9//LHzmJkzZ5r7NLA5REZGmvDmei7y5s1rXblyxXnfxIkTTRhyhCrX4KOhQQPaiRMn3NrToEEDa+DAgWkKPq5t/fnnn819UVFR5na7du2sJ598Msn/G9dzrT/z7bffdjtm2rRpVpEiRZy3NRDlzJnT6t+/v5U1a1Zr+vTpqZxVALiBoS7Yhn5G6zCMww8//CANGjQwwzo6bPX888+boZOrV6+m+jxbt26VZs2amaEa/b569eqZ+48dO5bs8VpcXapUKTNcoz9j+vTpbj/jwIED0q5dO/O4DkfdddddyT5ftWrVnF/rcJnSoTrX+xIPuekwkw4nOdSuXdsM5x0/fjxJO3ft2iXXr1+Xe+65xwydOS6rVq1yG3rzhGtbixQpYq4dbYuKijJDaK60Xa527NhhhrNc29GlSxdTu+U4d3q+3n33XRk1apQ8/fTTZsgRAG4l+JZHAAFCP3BLly7trEXRWppu3brJiBEjJG/evLJ27Vp56aWXTA2Ma1hwdeXKFWnUqJG5aIDR2hkNKHo7pdoZDUfbtm0ztTJLliyRwYMHm3ohrc/Ruh4NURqMJk+ebOpWtEalSpUqSZ5PZ6k5OAJc4vtup75FA1HmzJlNsNNrVxo80iK5tqalbdoWrelp1apVkse05sdh9erVpq36/1Prf4KDeUsDkDreJWALy5cvNz0affr0Mbf1w10/iN977z3JlOlGx6cWD7sKCQkxPSCu9u7da3qFRo4caQqi1ZYtW2758/UDuWHDhuYyZMgQE3i0TdpbtG/fPhN6tHBZaQBLL9pzcu3aNVMQrTZs2GBCjKPtrmrUqGFer/bMONpyJ1SsWFE2btzodp+2y5UWNet5KVu2bIrPM3v2bFNErYGyTZs2Mnz4cBOWACA1BB8EHJ2ZdOrUKfMhfvr0aTNrKjIy0vTwdOjQwRyjH6hxcXEyfvx40+Oybt06+eijj9yeR4ectOdh2bJlziEjHd7SQKTfpzOkdu/ebT5wU7Nw4UL55Zdf5JFHHpE8efLId999Z0JX+fLlzW2dyTVp0iQzJKS9RwMGDEi3c6G9RtqL9eabb5peEQ1dERERzrDnSoe42rdvb86RBkINQjpLTF+/Dl3prKn00LNnT6lbt64ZpmrevLl8//335v+RK+0V0/9fer51Bpy2V0Ocnu9//vOf8t///tf01ukwl86ImzJlijleZ7b95S9/SZd2AghQN2t9gICgBb36a62X4OBgM4upYcOG1qeffuos6HUYM2aMKZbVwthGjRpZn3/+ufm+8+fPO4/RWVJa8Kz3DxkyxNyns7N0hpHOdtKZRd988415XGdkJWfNmjWmiDhPnjzmZ1WrVs2aPXu28/GlS5eaomt9Pn1s5cqVyRYMuz7/ihUrkrQ1cTG2o9B78ODB5jVoUXOXLl2s6OjoFGd1xcbGmuP19WXJksWcn5YtW1o7d+5M8Zzfqq3aRr1P2+ygM+OKFy9uzofOYnv33XeTFJLrDDadYabHaHF4rVq1rEmTJlkJCQmm+Fn/n+nXDq+88opVpkyZFIvMAUAF6X+8Hb4ApD9dx+fChQu3vZ0EAAQSZnUBAADbIPgAAADbYKgLAADYBj0+AADANgg+AADANgg+AADANgg+AADANgg+AADANgg+AADANgg+AADANgg+AADANgg+AABA7OJ/1pr1c1byaSUAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 600x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAj4AAAEiCAYAAAD5ztjRAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjAsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvlHJYcgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAOGxJREFUeJzt3QmczdX/+PH3MGYGw9izSwjZsqRQURSSDGnxVZaibFn6KlS2fEUJhb76UtGiLEWLf4rsZN8iaxpLZas01pkx5vN/vA/3/u6dzR0zc7fP6/l4fNy5n8/n3ns+587c+3bO+5wTYlmWJQAAADaQw9cFAAAA8BYCHwAAYBsEPgAAwDYIfAAAgG0Q+AAAANsg8AEAALZB4AMAAGyDwAcAANgGgQ8AALANAh/YyvLly+Xhhx+WUqVKSVhYmBQsWFAqV64sjzzyiEyZMkViY2Od5954440SEhLilXIdPHhQ2rZtK0WKFJEcOXKY112xYoU5dvnyZRk2bJhUqFDBlFmPdenSxRxr0qSJuX/o0CGv18/1GjFihCnzzJkz3fbrNbleNwBkh9BseVbAD7366qsyfPhw83PVqlXl9ttvl1y5csm+fftk/vz58vnnn0u9evXkjjvu8Gq5kpKSpH379rJ9+3bz2pUqVTLBT/Hixc3xt99+W0aNGiUlS5aUdu3aSUREhNx55522qR8NQA8fPiysrgMgKxD4wBa2bNliWhr0i3zu3LkSHR3tdvz48ePyySefSIECBbxeNm2t0aDnrrvuklWrVqU4/uWXX5rb1atXy0033RS09TNmzBgZPHiwlC1bNtteAwAIfGAL2mKhLQaPPvpoii91pa0rAwcO9EnZfvvtN3ObVlBzrePBUj8lSpQwGwBkJ3J8YAunTp0yt0WLFr2ux7/33ntSs2ZNyZ07twkCnn32Wfnnn39SnJdeXpDmrrjm5yi937hxY/Pzhx9+aO7rprk7jpyXmJgY57mOzZOcnr///luGDBkit9xyiyl3VFSU3HvvvbJw4cIsqx/H9WrQpF1y+lraFac5Qn379k21jtKSPMfHUV/azZX8+vV1AeB60OIDWyhTpoy5/eKLL0wwUKxYMY8f++KLL5ovdQ1GKlasKGvXrpVp06bJnj17ZOXKlZlKgO7cubPpRvr+++9N8rIjd6dKlSom0Vlpbs358+fNuQ6RkZHpPu/+/fulWbNmcvToURMkNG/eXM6ePSvr16+X1q1by7hx49xacDJTP+q5554zdaJ1VKNGDVMvkydPNrfaRZc/f37JKA0w9ZpTu35H3QBAhlmADRw8eNDKnTu3Zsda+fLlszp37mxNnz7d2rp1q5WYmJjqY8qVK2fOL168uLV3717n/lOnTlkVK1Y0x5YuXZrqY1KzfPlyc0xf25P9njxn48aNzbGYmBjnPr2eGjVqmP1vvPGGdfnyZeexAwcOWOXLl7dy5sxp7dy5M1P141q2/PnzW5s3b3buP3v2rHXvvfeaY/369XN7zPDhw83+GTNmuO3X19T9Wh+eXj8AZBRdXbAFzY/55ptvTMuGtnxot1L37t2lTp06pvWgV69ecuzYsVQfqyOqdEi3g57fo0cP83Nqyci+pte5c+dOMyz9hRdeMCPEHLTFavz48WaI/PTp07OkflSfPn2kbt26bi1S2uKjrWHvv/++xMXFZeMVA4DnCHxgG02bNpVffvnFJPJq4KJf6qGhoSYPZerUqXLrrbeaodvJ3X///Sn23XzzzeY2vWDAVxYvXmxudeh7anT0mNq4cWOW1I96/PHHU+zTfJ9atWrJuXPnZNu2bVlwZQCQeQQ+sBWdlE8nCtQvch3CrUm9+rNO1Hfy5EnTcpFc6dKlU+zLly+fuY2Pjxd/40h87tixo1tCsGNzJDD/+eefWVI/qly5cqnudyQh//HHH1l4hQBw/Uhuhq3pvDTauqGTA7Zp08bMXHzhwgXJkyeP8xzXrqLMTlToDY7XadGihdxwww1pnudJgrAn9QMAgYTABxAxw7yV5r5o1871frFri4nS7p3kI690hJU3OFqounXrZvJ8vFE/OuRcR3Ml5xiKroETAPgDurpgC9da7kBzWxyBS2aGSjsm4NPh5MktWbJEvOG+++4ztwsWLPBa/ehsz8nt3bvXzEitAaDmB10vRzCZmJh43c8BAA4EPrCFoUOHmhFOuhhocr///ruZkFA99NBDzi/a6+GYjFCXX9DWEYfPPvvMbN6grTyaWDxr1iwzIi15HpIGOToXkW5ZVT86gss1gVm7w3RuH32trl27mgkUr5ejtSitxGoAyAi6umAL2vWkkxC++eabZkSWY4ZhXQ5iw4YNcunSJTPU+6233srU6/Tu3VveffddM+mevobO9nzgwAHZtWuX9OvXTyZOnCjZTUdi6fpeOmmhruquq6prOXRSQk1o1lYYTVTWsjRq1ChL6ueJJ54wi5pql5jOEK3D/HVixmrVqpngKzM02NKJEHXU2T333CN58+Y1rU5jx47N1PMCsCcCH9jCK6+8YlYW1xmSd+zYYWYTjo2NNTMK169f3yTu6lw1+qWaGZpMrF/62nqiX9baWqLz22g3l46o8kbgo3SFd22B0aBHh6frjM3aVaSzIdeuXdsEE7ouV1bVz6RJk6R8+fJmaQ9dYqNQoUImCNSgRwOhzNClL06fPm1azHRmaQ3CdBQZgQ+A6xGisxhe1yMB2J4OV9cEZj5GAAQKcnwAAIBt0NUFAPBLOkBAuzaBa8mVK5fkzJlTPEHgAwDwK9p1qsnxOmcU4CmdcFXzGDWfMj3k+AAA/IqugadBj45E1Mkyr/VFBnuzLMtMoaGjVTX4ccynlhZafAAAfsMxO7gGPYULF/Z1cRAgHHOFafCjvzvpdXuR3AwA8BuOnB7Wg0NGOX5nrpUXRouPh4s+6urSuiI3Ta4AkPmuibNnz5pZudNaBJjPWmSUp78zBD4e0KCnTJkyvi4GAAQVXbjXsagu3OfH6t+/v9nS+5LX9fiio6PF244fPy5PPvmk/Pjjj2Y0lSdJ6CNGjDAzyuvM8Wnp0qWLeS49LzsR+HhAW3ocf6Q6ky0A4PqdOXPG/GfS8dnqiaTYvyXpwjnxlhx5IiVHVCGPz0/rS3vFihVmqRWdfVwTbwM94FI6A70moGsQk9mZ2X2BwCcDzWca9BD4AIB3uyY06ImdOlzkcqJ4Tc5Qieo5MkPBj10cPHjQLMWjS+MEIpKbAQB+zbT0eDPoUZcTs62FSdec0wV8w8PDTSvL+PHj0z1fFzq+++67zcLBuoCwrv2X3M6dO80iwTq6SUfDPfPMM2bxYYcmTZqkaMmJjo42LVWO47r8zIABA0xAmlZQquXV8n/00UfmHMfjjxw5Ytb0i4yMNA0EuhbgiRMn0h299/zzz5tWMC3viy++6LWlbwh8AADwki1btpig4PHHHzfBiua+DB06VGbOnJnm4Jp27dpJWFiYbNiwQd59910ZNGiQ2znnz5+X5s2bS8GCBWXTpk0yb948+eGHH6RPnz4el2v+/Pkm3+rVV1813Vi6pUafv0WLFuYa9Jy3337blFGDnr///tsszqyB2a+//iqPPfZYmq+nwZ5e8wcffCBr1qwxj9WcJW+gqwsAgCywcOFC0+KRvGXD1YQJE6Rp06Ym2FE333yz7N69W8aNG+dsPXGlAczevXvl+++/N6Pg1GuvvSYtW7Z0nvPpp59KXFycaYXJmzev2TdlyhRp3bq1vP7663LDDTdcs+yFChUyc99o3pXOfpyWokWLmpYqbVlynKeBjgZxMTExzoFAWhZt1dJA6bbbbkvxPG+99ZYMGTLEBHVKAzq9Rm+gxQcAgCygScya8Ou6vffee27n7NmzRxo1auS2T+9rd1byIMlxvgYTjqBHNWjQIMU5tWrVcgY9jufUlph9+/ZJdttztYyuo5+1S067sfRYcrGxsaa16Pbbb3fuCw0NlXr16mV7Wc1reeVVAAAIchp4VKxY0W3fb7/9Jv5A50tKnkNzyaYLwNLiAwCAl1StWlXWrl3rtk/va5dXasss6Pk6lYprzs369etTnLNjxw6T6+P6nBrsVK5c2dlF5focly9fll27drk9j+YRpdbq5Mk1aRl1c9DuOx3ery0/yekQeF1PS3OWHBITE03+kzcQ+AAA4CX//ve/ZenSpTJq1CjZv3+/fPjhhyYfZ+DAgame36xZMxMUde7c2QQ3q1evlpdfftntnI4dO5oRX3qOBjPLly+X5557zkwy6Mjv0RFf/+///T+zac5Qz549U0w8qCO2Vq1aJb///rv8+eefHl+TlrFGjRqmHFu3bpWNGzdKp06dpHHjxml2X/Xr10/Gjh1r5j3S8vTq1cujiRCzAoEPAABeUqdOHZk7d67Mnj1bqlevLsOGDTMjqVJLbFbaaqOjnS5evCj169eXbt26yejRo1OsUaWJwToyShOJ27dvbxKoNaByeOqpp0xg1OlqQHLTTTeZnCRXWo5Dhw5JhQoVTAuRp3RY+1dffWVGlemwew2E9PnnzJmTbgCogZmWSXOWNKm6bdu24g0hlrcGzgf4LKPaNKcJWUxgCADZ95mqo5N0dFD58uVNK4ZiAkN4IrXfndSQ3JwB/SeelrCIjPd/qoFH+ohcnQ8qLCnO3KY2PdS4MpPd7ieERGg4LXZi6kqFiEwqNc6r9eH6PmX3a7teZ/LXC6brDIRyZNl7Zo75tsz+XMaIsCu38XFnMvQ4DT40CPHnJSsQOAh8vCTcihPxoG0tIUdusTtTV8ryfn24vk/Z/dqu1+mN10vx2j54XX8tRyC8Z8FQxriEK7cJV28zQoMQAhFkBXJ8AACAbdgm8HnnnXdMxrr2++mkSZp1DgAA7MUWgY9mlutiaMOHDzdD7XSGS13X5OTJk74uGgAA8CJbBD66Nkr37t2la9euZjIlXRNEh//p4mgAAMA+gj7wSUhIMLNB6rwCrvMi6P1169al+pj4+Hgz3NJ1AwAAgS/oAx+dfVKn4E6+Oq3eP378eKqPGTNmjJljwrG5LrwGAAACV9AHPtdjyJAhZmItx+a6/ggAAAhcQR/4FClSxCz8duLECbf9er948eKpPiY8PNzMJuq6AQDgTToS+a233vL4/BUrVpjlI7yx5tWXX35pVqLX79f+/ftn2fVo+fW5s1PQT2Coq83WrVvXLAoXHR1t9iUlJZn7ffpcneEUAODXTvwZL7FnvLdkRVT+ULmhSLhH5+qXdXp0RPGIESMyXIZNmzZJ3rx5PT6/YcOGZgV2TdHIbs8++6wZMNS3b1+zzlYgCfrAR+lQdl0ITVeJ1UXeNOI8f/68edMAAP4f9HTqt0MSLnlvacmwXCHy0du1PAp+NNhwnT5FFx7dt2+fc19kZKTzZ10eU/NOQ0Ov/fWbkYVCTZnDwtLsychK586dM9PB6LQwJUuWlEAT9F1d6rHHHpM333zT/DLeeuutsn37dvnuu+9SJDwDAPyPtvR4M+hR+nqetjBpsOHYtLVFW4Ac9/fu3WtaRBYtWmR6HzSVYs2aNXLw4EFp06aN+R7SwEhXVf/hhx/S7RrS533vvffMKuY6JUulSpXk66+/TrOra+bMmVKgQAGzcnvVqlXN67Ro0cItUEtMTDStNnpe4cKFZdCgQaahwNFDkpy+hqOF59577zWvp/vUF198IdWqVTPXqGUfP358uvV24MABs5q7TiysU80sWbLE/T1ISDA9MyVKlDDnlCtXzgw+yixbBD5KK+/w4cNmqPqGDRvM7M3eFB8SIfE5rmz655vWn3BY0kW3TSzv/rH7A1NXV+vL2/Xh+j5l92u7Xmfy1wum6wyEcmTZe+YHZfbnMuoipbqFX12s1E4GDx4sY8eOlT179kjNmjVNq8kDDzxg0i62bdtmApLWrVvLkSNH0n2ekSNHyqOPPio//fSTeXzHjh3l77//TvP8CxcumP/4f/zxx7Jq1Srz/AMHDnQef/3112XWrFkyY8YMWbt2rZm+Jb0cG+1Oc7RmaaCjQZTu02ljtFyPP/647Ny503TtDR061ARfqdGUk3bt2plWKv1O1vn1NOhyNWnSJBPYzZ0717ymllMDqsyyRVdXVnlrQMFMJDrP8uisSdf57MFllg/rw5uv7f47Mck2deyP5QiE9yx4ynjmTE559yWxlVdffVXuu+8+5/1ChQqZVQQcRo0aJQsWLDBf9Onln3bp0kU6dOhgfn7ttddMcKBLMGnglJpLly6ZoKJChQrmfp8+fUxZHCZPnmxGMmsrkpoyZYp8++23ab6+BirFihVzXoOja00nCm7atKkJdtTNN98su3fvlnHjxpkyJ6etW9oapq1Rju4yvZ6WLVs6z9EgTVu17rzzTtOypC0+WcE2LT4AAPiK5pi60hYfbXnRLijtZtJuKG0NulaLj7YWOWjis/5nPL3ll7RLzBH0qBIlSjjP1+ladISz5r466Cgt7ZLLKC17o0aN3Pbpfe3O0pym1M7XOfJcc4QaNGjgdo4GTJqaUrlyZdMdt3jxYskKBD4AAGSz5KOzNOjRFh5t5Vi9erX5gq9Ro4bJa0lPrly53O5rS4h2G2XkfMsPumQ9UadOHYmJiTGtYRcvXjRdae3bt8/08xL4AADgZZpPoy0a2sWkAY92GR06dMirZYiKijLJ1Tps3kFbZ3Qx74zSliu9Jld6X7u8tBUptfN1cmDXROv169enOE9btHSA0vTp082IOc0rSi+nyRPk+AAA4GWauzJ//nyT0KytMJobk17LTXZ57rnnzEgpnYywSpUqJufn9OnT15ybKLl///vfZmSats5ooKJrYWq+0H//+99Uz9f1MjUo0hFkmgekSdUvv/yy2zmaN6Rdc7Vr1zZrbM6bN88EiNo1mBm0+AAA4GX6pV6wYEEzIkqDH50TR7t2vG3QoEEmWbpTp04mx0ZzjbQsOnw8I7TsOvpq9uzZUr16dTN9jCZRp5bYrDSQ0a4+7cLSHKNu3brJ6NGj3c7RYfNvvPGGyY/SoEpbxDTxWh+bGSFWoHT2+ZBGotokqIlgLF8BANn3mRoXF2fyOsqXL+/88vX3CQyDSVJSkumG0nwabb0JJKn97qSGri4AgF/T4EODEH9dsiKQHT582IyWaty4sZnnTrunNHj417/+JcGKwAcA4Pc0CLFDIOJtOXLkMJMM6igz7QDSbiqdY0dbfYIVgQ8AADZVpkyZFKOxgh3JzQAAwDYIfAAAgG0Q+AAA/A4DjpFdvzMEPgAAv+FYYkFXFQcywvE7k3yZjuRIbs6A/hNPS1hEysXWHAYeubqi7tUJLyeVGuc8lhASoYukpP24VB5zrcdlNX8pR1aUPywpztw6SjyuzOSAuhZk8G/LHPP+expofzOefkYlP0//nlL7W0r+uGuJCLtyGx93Js1zdHkDnZnXsZCmLrKZ0VmEYb+WngsXLpjfGf3dSW2JDFcEPlko3LryZStXW9sScuT2/HEZfEx28JdyZEX5kwvE60Hm/7ayW6D9zXhaj8nPc5WZ64y7uv7mNdbhNMsSqPRWHQeS06DH8bsjdg98Vq1aZdYC2bJli1kQTafJjo6O9nWxAACp0BYeXaOpWLFicunSJV8XBwFAu7eu1dJjq8Dn/PnzUqtWLXnqqaekXbt2vi4OAMAD+kXm6ZcZ4ClbBD4tW7Y0GwAAsDdGdQEAANuwRYtPRulCbbq5riQMAAACHy0+qRgzZoxERUU5N13LBAAABD4Cn1QMGTJEYmNjndvRo0d9XSQAAJAF6OpKRXh4uNkAAEBwsUXgc+7cOfnll1+c92NiYmT79u1SqFAhKVu2rE/LBgAAvMcWgc/mzZvlnnvucd5//vnnzW3nzp1l5syZPiwZAADwJlsEPk2aNGGlXwAAYI/Ax1vinYslXrkJS7ro0UJ+5nGpPOZaj8tq/lKOrCm/+yKlnr4XCNC/LXPM++9poP3NePoZlfK8/1ukNDPX6VikNCQpM1cBZE6IRVPINek8PjqsXUd45c+f39fFAYCAxmcqfInh7AAAwDYIfAAAgG0Q+AAAANsg8AEAALZB4AMAAGyDwAcAANgGgQ8AALCN65rA8MCBA7J8+XI5efKkJCW5z0Q1bNiwrCobAACAbwOf6dOnS8+ePaVIkSJSvHhxCXGZsVN/JvABAABBE/j85z//kdGjR8ugQYOyp0QAAAD+kuNz+vRpeeSRR7KnNAAAAP4U+GjQs3jx4uwpDQAAgD91dVWsWFGGDh0q69evlxo1akiuXLncjvft21eCVf+Jp+Wlkz2v3Lma2jSp1Di/WCXa1cAjfa6WI3vKmN3PH2iyqj7M8wRhnSavH9eVvseVmRw01+kvfPn36fra+j67FOPKzxG5ze3Z+EvZ8vpAtgQ+06ZNk8jISFm5cqXZXGlyczAHPircuvLHLFfXtE/IceUP2U5lDIQ68Kasqg/zPEFYp8nrx1UwXae/8OXfZ3rvtdkdd/HKbQKBDwIo8ImJicmekgAAAPjzBIaWZZnNn40ZM0Zuu+02yZcvnxQrVkyio6Nl3759vi4WAAAIlMDno48+Mvk9uXPnNlvNmjXl448/Fn+k3XG9e/c2OUlLliyRS5cuyf333y/nz5/3ddEAAIC/d3VNmDDBJDf36dNHGjVqZPatWbNGevToIX/++acMGDBA/Ml3333ndn/mzJmm5WfLli1y9913+6xcAAAgAAKfyZMny9SpU6VTp07OfQ899JBUq1ZNRowY4XeBT3KxsbHmtlChQmmeEx8fbzaHM2fOeKVsAADAz7q6jh07Jg0bNkyxX/fpMX+m64r179/ftFRVr1493bygqKgo51amTBmvlhMAAPhJ4KPz+MydOzfF/jlz5kilSpXEn2muz65du2T27NnpnjdkyBDTMuTYjh496rUyAgAAP+rqGjlypDz22GOyatUqZ47P2rVrZenSpakGRP5Cc5IWLlxoyl26dOl0zw0PDzcbAACweYvPww8/LBs2bDCrs3/55Zdm0583btwobdu2FX+jw+016FmwYIEsW7ZMypcv7+siAQCAQGnxUXXr1pVPPvlEAoF2b3366afy1Vdfmbl8jh8/bvZr7o4OxQcAAPbhUeCjo5ry58/v/Dk9jvP8hY5AU02aNHHbP2PGDOnSpYuPSgUAAPw28ClYsKAZsaXz3xQoUMCsyZVal5Luv3z5sviTrJ5ZOt65yN+Vm7CkK2vP+NMii9ldxkCoA2/KqvowzxOEdZqyfv5vkdJguk5/4cu/T9fXTm+R0pCQ6+psALJEiOVBZKCzH2sic2hoaIqFSZNr3LixBBtt5dKuMR3h5W8tWgAQaPhMhS95FHa7BjOaHKzz2iRv9dH4iWHfAAAgqEZ1aeBz6tSpFPv//vtvRkwBAIDgCnwcuTzJnTt3TiIirvbvAgAA+CGPM8yef/55c6tBjy5SmidPHucxTWjWuX1uvfXW7CklAACANwOfbdu2OVt8du7cKWFhYc5j+nOtWrVk4MCBWVEmAAAA3wY+y5cvN7ddu3aVt99+m0x8AAAQcDI8mYJO/AcAABCIrmsWqc2bN5sFSY8cOSIJCQlux+bPn59VZQMAAPDtqK7Zs2dLw4YNZc+ePWbhz0uXLsnPP/9sFgDVCakAAACCJvB57bXXZOLEifLNN9+YpGbN99m7d688+uijUrZs2ewpJQAAgC8Cn4MHD0qrVq3Mzxr4nD9/3gxxHzBggEybNi0rygQAAOAfgY8uWHr27Fnzc6lSpWTXrl3m53/++UcuXLiQ9SUEAADwVXLz3XffLUuWLJEaNWrII488Iv369TP5PbqvadOmWVUuAAAA3wc+U6ZMkbi4OPPzyy+/LLly5ZIff/xRHn74YXnllVckmPWfeFrCIi6LPxp4pI/I1ZVEJpUa53YsISRCp9z2XjlUSPrlcD1PuZ5rzjPHQjw+D9fPX+s4K36n/eXvwl/ez76/vZDmsayuj+TlUBFhV+6cjb+UZa8DZGvgk5iYKAsXLpTmzZub+zly5JDBgwdn+EWR9cKtOBHrys8JOXL7thzKSr8crueptM719DxcP3+t46z4nfaXvwt/eT+9+V4nfy3zo2NXAoEPAiTHJzQ0VHr06OFs8QkEU6dOlZo1a5qZpnVr0KCBLFq0yNfFAgAAgZDcXL9+fdm+fbsEitKlS8vYsWNly5YtZuLFe++9V9q0aWPmHgIAAPaS4RyfXr16mZXajx49KnXr1pW8efO6HdfWFX/SunVrt/ujR482rUDr16+XatWq+axcAAAgAAKfxx9/3Nz27dvXuU/n8dFV2/X28mX/TP5VWrZ58+aZuYe0ywsAANhLhgOfmJgYCTQ7d+40gY7mJkVGRpqlNm655ZY0z4+Pjzebw5kzZ7xUUgAA4FeBT7ly5STQVK5c2eQlxcbGyueffy6dO3eWlStXphn8jBkzRkaOHOn1cgIAAD9LblYff/yxNGrUSEqWLCmHDx82+9566y356quvxB/p0hoVK1Y0OUka1NSqVcusMZaWIUOGmCDJsWk+EwAAsGHgo4nBmtz8wAMPmGUqHDk9BQoUMMFPIEhKSnLrykouPDzcOfzdsQEAABsGPpMnT5bp06ebWZtz5szp3F+vXj2TS+NvtPVm1apVcujQIVM+vb9ixQrp2LGjr4sGAAACIbm5du3aqbaS6Ggpf3Py5Enp1KmTHDt2TKKiosxw+++//17uu+8+XxcNAAD4e+BTvnx5kyicPMn5u+++k6pVq4q/ef/9931dBAAAEKiBj+b39O7d2wwN17l7Nm7cKJ999plJGn7vvfeyp5S4pnizwOCVn8OSLvpsMUZTDhWSfjlcz1Ou57oujOnpebh+/lrHWfE77S9/F/7yfl7zvc7C+kj+Wq6LlIaEZPirB8gyIZZGLxk0a9YsGTFihBw8eNDc19FdOvz76aeflmCk8/hoN5mO8CLRGQAyh89UBFzg43DhwgU5d+6cFCtWTIIZf6QAkHX4TEVAjerSRT51GLvKkyePM+jRX2Q9BgAAEDSBjw4FT0hISLFfc35Wr16dVeUCAADIch5nmP3000/On3fv3i3Hjx933tdJDHVUV6lSpbK+hAAAAN4OfG699Vaz+rpuqXVp5c6d20xuCAAAEPCBj05cqHnQN910kxnCXrRoUbe1sDTXx3UmZwAAgIANfBwTFuo6VwAAAEEd+Hz99dcenffQQw9lpjwAAAC+D3yio6OveY7m/zhWawcAAAjYwIcuLgAAYLt5fAAAAAIVgQ8AALANlsjNgP4TT0tYRPDmMA080se5kvKkUuPcjnm6crN5DhVy/c/hTa7lVWFJcc7FpMeVmez35Q/E9zOtcvX97QXnayU/5stV4tP7ffHHMvry9+VaIsKu3MbHncn0cwHXi8AHTuFWnMjVJWsTcuS+/udQ1vU/hze5lje5QCh/IL6faZUr+Xvhj/UfCGX05e/LtcRdXe0olVWPAP/r6gqW0Vpjx441o8/69+/v66IAAAB/DXx0Ha7BgwfL/v37JVBt2rRJ/ve//0nNmjV9XRQAAODPgU/v3r3l888/l6pVq8pdd90lM2fOlAsXLkigOHfunHTs2FGmT58uBQsW9HVxAACAPwc+Q4cOlV9++UWWLl1q1uvq06ePlChRQrp37y4bNmwQf6eBW6tWraRZs2a+LgoAAAiU4exNmjSRDz/8UI4fPy7jx4+XPXv2SIMGDaRatWoyYcIE8UezZ8+WrVu3ypgxYzw6Pz4+Xs6cOeO2AQAAG8/jExkZKd26dZM1a9bIN998YwKhF164OhzVjxw9elT69esns2bNkoiIq8NNr0EDpKioKOdWpkyZbC8nAADw48BH83s0z6dx48ZmYdLChQvL6NGjxd9s2bJFTp48KXXq1JHQ0FCzrVy5UiZNmmR+Tm202pAhQyQ2Nta5afAEAABsOI/Pjz/+KB988IHMmzdPEhMTpX379jJq1Ci5++67xR81bdpUdu7c6bava9euUqVKFRk0aJDkzJkzxWPCw8PNBgAAbBr4vPHGGzJjxgwznL1evXoybtw46dChg+TLl0/8mZavevXqbvvy5s1rWqiS7wcAAMHN48BHA50nnnjCtPQQMAAAgKAOfP744w/JlSuXBIMVK1b4uggAAMCfk5tXr14tt9xyS6pDuzUBWIez6zkIXPEhERKf48oWlnTRbRPL8vw5QjL3HN7kWl7dXEsYCOUPxPczrXIlfy9SvJYf1H8glNGXvy+eLFKqW/jVxUoBXwixLM9+o3Xk1j333CMDBgxI9biOklq+fLksWLBAgo0GezqsXQO8/Pnz+7o4ABDQ+ExFQLT47NixQ1q0aJHm8fvvv98MHQcAAAj4wOfEiRPp5vjonDinTp3KqnIBAAD4dnX2Xbt2pXn8p59+Mmt3AQAABHzg88ADD5iFSuPi4lIcu3jxogwfPlwefPDBrC4fAACA95ObtatLl33QmY51ZfbKlSub/Xv37pV33nnHLP2gC4HecMMNEmxIxAOArMNnKgJiHh8NaHS5ip49e5q1rBzxUkhIiDRv3twEP8EY9AAAAJuu1VWuXDn59ttv5fTp0/LLL7+Y4KdSpUpSsGDB7CshAACArxYpVRro3HbbbVlVBgAAAP9KbgYAAAh0BD4AAMA2CHwAAIBtEPgAAADbuK7kZruat+N3yROZcnX67NQudpLz51BJMLchXi0BUvPdjB+u/hQiiZcSfVwaIDCE5s5rbi8kXPZ1UWBjBD5+LtfVYAf+JfESH9xARiVePH/llr8f+JBfdnUdOnTITIy4fft2XxcFAAAEEb8MfAAAALIDgQ8AALANnwY+SUlJ8sYbb0jFihUlPDxcypYtK6NHj05xni6A+vTTT0v58uUld+7cZoHUt99+2+2cFStWSP369SVv3rxSoEABadSokRw+fNgc27Fjh9xzzz2SL18+syBe3bp1ZfPmzV67TgAA4B98mtysi51Onz5dJk6cKHfeeaccO3bMrPaeWoBUunRpmTdvnhQuXNgslvrMM89IiRIl5NFHH5XExESJjo6W7t27y2effSYJCQmyceNGkyekOnbsKLVr15apU6ea1eU1dyhXrlxplis+Pt5srisJAwCAwOezwOfs2bOm1WbKlCnSuXNns69ChQomANLkZlcapIwcOdJ5X1t+1q1bJ3PnzjWBjwYmsbGx8uCDD5rnUFWrVnWef+TIEXnhhRekSpUq5r4urJqeMWPGuL0eAAAIDj7r6tqzZ49pVWnatKlH57/zzjumi6po0aISGRkp06ZNMwGNKlSokHTp0kWaN28urVu3NgGVth45PP/889KtWzdp1qyZjB07Vg4ePHjNligNpBzb0aNHM3m1AADA1oGP5up4avbs2TJw4ECT57N48WLTVdW1a1fTpeUwY8YM0wrUsGFDmTNnjtx8882yfv16c2zEiBHy888/S6tWrWTZsmVyyy23yIIFC9J8Pc030lwg1w0AAAQ+nwU+2t2kwc/SpUuvee7atWtNQNOrVy+Tq6PJ0Km12ugxba3RHKDq1avLp59+6jymgdCAAQNM4NSuXTsTKAEAAHvxWeATEREhgwYNkhdffFE++ugjE8hoC83777+fapCko7C+//572b9/vwwdOlQ2bdrkPB4TE2MCHm3x0ZFcGtwcOHDA5PlcvHhR+vTpY0Z96TENovSxrjlAAADAHnw6qksDmNDQUBk2bJj88ccfZpRWjx49Upz37LPPyrZt2+Sxxx4zI7U6dOhgWn8WLVpkjufJk8eMBvvwww/lr7/+Ms/Tu3dv8zgd8aX7OnXqJCdOnJAiRYqYFh+SlwEAsJ8Qy7IsXxfC3+mosaioKHlv1W7JE5nPq6/NIqX+iUVKgcwtUvrIJ9vN4BFyKOFtBD4ZCHz4IwWAzOMzFb7EkhUAAMA2CHwAAIBtEPgAAADbIPABAAC2QeADAABsg8AHAADYBoEPAACwDQIfAABgGwQ+AADANgh8AACAbRD4AAAA2yDwAQAAtkHgAwAAbIPABwAA2AaBDwAAsA0CHwAAYBsEPgAAwDYIfAAAgG0Q+AAAANsg8AEAALZB4AMAAGyDwAcAANgGgQ8AALANAh8AAGAbBD4AAMA2CHwAAIBtEPgAAADbIPABAAC2QeADAABsg8AHAADYBoEPAACwDQIfAABgGwQ+AADANgh8AACAbRD4AAAA2yDwAQAAtkHgAwAAbIPABwAA2EaorwsQCCzLMrdnzpzxdVEAIOA5Pksdn62ANxH4eOCvv/4yt2XKlPF1UQAgqD5bo6KifF0M2AyBjwcKFSpkbo8cOcIf6dX/rWkQePToUcmfP7/YHfXhjvpIiTpxFxsbK2XLlnV+tgLeRODjgRw5rqRCadDDh9b/0bqgPv4P9eGO+kiJOkn9sxXwJn7rAACAbRD4AAAA2yDw8UB4eLgMHz7c3IL6SI76cEd9pESduKM+4EshFuMJAQCATdDiAwAAbIPABwAA2AaBDwAAsA0Cn2t455135MYbb5SIiAi5/fbbZePGjWIHY8aMkdtuu03y5csnxYoVk+joaNm3b5/bOXFxcdK7d28pXLiwREZGysMPPywnTpwQOxg7dqyEhIRI//79bV0fv//+uzzxxBPmmnPnzi01atSQzZs3O49rCuGwYcOkRIkS5nizZs3kwIEDEowuX74sQ4cOlfLly5trrVChgowaNcptWYZgro9Vq1ZJ69atpWTJkuZv48svv3Q77sm1//3339KxY0cz11GBAgXk6aeflnPnznn5ShDsCHzSMWfOHHn++efN6IOtW7dKrVq1pHnz5nLy5EkJditXrjRf4uvXr5clS5bIpUuX5P7775fz5887zxkwYIB88803Mm/ePHP+H3/8Ie3atZNgt2nTJvnf//4nNWvWdNtvt/o4ffq0NGrUSHLlyiWLFi2S3bt3y/jx46VgwYLOc9544w2ZNGmSvPvuu7JhwwbJmzev+RvSIDHYvP766zJ16lSZMmWK7Nmzx9zX6588ebIt6kM/G/QzUv+zmBpPrl2Dnp9//tl85ixcuNAEU88884wXrwK2oKO6kLr69etbvXv3dt6/fPmyVbJkSWvMmDGW3Zw8eVL/22qtXLnS3P/nn3+sXLlyWfPmzXOes2fPHnPOunXrrGB19uxZq1KlStaSJUusxo0bW/369bNtfQwaNMi688470zyelJRkFS9e3Bo3bpxzn9ZTeHi49dlnn1nBplWrVtZTTz3ltq9du3ZWx44dbVcf+nu/YMEC531Prn337t3mcZs2bXKes2jRIiskJMT6/fffvXwFCGa0+KQhISFBtmzZYppjXadX1/vr1q0TO66toxxr62jdaCuQa/1UqVLFrL8TzPWjrWCtWrVyu2671sfXX38t9erVk0ceecR0h9auXVumT5/uPB4TEyPHjx93qxNd9kW7jIOxTho2bChLly6V/fv3m/s7duyQNWvWSMuWLW1ZH648uXa91e4t/Z1y0PP1c1dbiICswlpdafjzzz9Nn/0NN9zgtl/v7927V+wkKSnJ5LJot0b16tXNPv0QCwsLMx9UyetHjwWj2bNnmy5P7epKzo718euvv5quHe0Ofumll0y99O3b19RD586dnded2t9QMNbJ4MGDzWKkGvDmzJnTfH6MHj3adN8ou9WHK0+uXW81gHYVGhpq/rMV7PUD7yLwgUetHLt27TL/e7UrXVW7X79+JvdAE91xJSDW/52/9tpr5r62+OjvieZwaOBjN3PnzpVZs2bJp59+KtWqVZPt27eb/zBosq8d6wPwV3R1paFIkSLmf23JR+Xo/eLFi4td9OnTxyQZLl++XEqXLu3cr3Wg3YH//POPLepHu7I0qb1OnTrmf6G6aQKzJmvqz/o/VzvVh9LRObfccovbvqpVq8qRI0fMz47rtsvf0AsvvGBafR5//HEzuu3JJ580Ce86QtKO9eHKk2vX2+QDRxITE81Ir2CvH3gXgU8atLm+bt26ps/e9X+4er9BgwYS7DQ/UYOeBQsWyLJly8wQXVdaNzqax7V+dLi7fukFY/00bdpUdu7caf4X79i0tUO7MRw/26k+lHZ9Jp/iQPNbypUrZ37W3xn9wnKtE+0K0nyNYKyTCxcumHwUV/qfJ/3csGN9uPLk2vVW/+Og/8lw0M8erT/NBQKyjK+zq/3Z7NmzzaiDmTNnmhEHzzzzjFWgQAHr+PHjVrDr2bOnFRUVZa1YscI6duyYc7tw4YLznB49elhly5a1li1bZm3evNlq0KCB2ezCdVSXHetj48aNVmhoqDV69GjrwIED1qxZs6w8efJYn3zyifOcsWPHmr+Zr776yvrpp5+sNm3aWOXLl7cuXrxoBZvOnTtbpUqVshYuXGjFxMRY8+fPt4oUKWK9+OKLtqgPHfG4bds2s+lXy4QJE8zPhw8f9vjaW7RoYdWuXdvasGGDtWbNGjOCskOHDj68KgQjAp9rmDx5svkyCwsLM8Pb169fb9mBfnClts2YMcN5jn5g9erVyypYsKD5wmvbtq0Jjuwa+NixPr755hurevXq5j8IVapUsaZNm+Z2XIcxDx061LrhhhvMOU2bNrX27dtnBaMzZ86Y3wf9vIiIiLBuuukm6+WXX7bi4+NtUR/Lly9P9TNDA0JPr/2vv/4ygU5kZKSVP39+q2vXriagArISq7MDAADbIMcHAADYBoEPAACwDQIfAABgGwQ+AADANgh8AACAbRD4AAAA2yDwAQAAtkHgAwAAbIPAB7CxJk2amBXEMyMkJES+/PLLTD3HzJkzpUCBApl6DgDwBIEPgkqXLl3MF7Fuumiorpp+3333yQcffOBcLNJTfBl75tixY9KyZUtfFwMAPELgg6DTokUL82V86NAhWbRokdxzzz3Sr18/efDBByUxMdHXxQs6uup2eHi4r4sBAB4h8EHQ0S9h/TIuVaqU1KlTR1566SX56quvTBCkrTgOEyZMkBo1akjevHmlTJky0qtXLzl37pw5tmLFCunatavExsY6W5BGjBhhjn388cdSr149yZcvn3mdf/3rX3Ly5Ml0y/Tf//5XKlWqJBEREaYVqn379s5j3333ndx5552mdalw4cImQDt48KDzuAZw+vpz586Vu+66S3Lnzi233Xab7N+/XzZt2mTKEhkZaVpdTp065db6FR0dLSNHjpSiRYtK/vz5pUePHpKQkJBmOePj42XgwIGm7rRebr/9dlMXnnZ1Oco6f/58E3DmyZNHatWqJevWrXN7jL4PZcuWNcfbtm0rf/31V4rn1fdM3z+ts5tuuslchyNwffXVV6VkyZJuj2vVqpV5zYy27AGwmSxd8hTwMV0Juk2bNqkeq1WrltWyZUvn/YkTJ1rLli2zYmJirKVLl1qVK1e2evbsaY7pitpvvfWWWSFaV1jXzbFK9Pvvv299++231sGDB61169ZZDRo0cHve5DZt2mTlzJnT+vTTT61Dhw5ZW7dutd5++23n8c8//9z64osvrAMHDljbtm2zWrdubdWoUcO6fPmyOa7l0z9VXf38u+++s3bv3m3dcccdVt26da0mTZpYa9asMc9ZsWJFq0ePHm51oatcP/bYY9auXbushQsXWkWLFrVeeumlNFeY79atm9WwYUNr1apV1i+//GKNGzfOrKS9f//+NK9Py7ZgwYIUZdXX09W327dvb5UrV866dOmSOWf9+vVWjhw5rNdff90c17ooUKCAFRUV5XxOfX2t+5kzZ5p6Xrx4sXXjjTdaI0aMMMcTExNNvUdHR5v7U6ZMMc9x+PDhNMsJAIrAB7YJfDQAqFq1apqPnTdvnlW4cGHn/RkzZrh9GacX2OiXvSMwSk6DGv0SP3PmjEfXcOrUKfN8O3fudAsm3nvvPec5n332mdmnAZvDmDFjTPDmWheFChWyzp8/79w3depUEww5girXwEeDBg3Qfv/9d7fyNG3a1BoyZEiGAh/Xsv78889m3549e8z9Dh06WA888ECK98a1rvU1X3vtNbdzPv74Y6tEiRLO+xoQ5cuXzxo0aJCVO3dua9asWenUKgBcQVcXbEO/o7UbxuGHH36Qpk2bmm4d7bZ68sknTdfJhQsX0n2eLVu2SOvWrU1XjT6ucePGZv+RI0dSPV+Tq8uVK2e6a/Q1Zs2a5fYaBw4ckA4dOpjj2h114403pvp8NWvWdP6s3WVKu+pc9yXvctNuJu1OcmjQoIHpzjt69GiKcu7cuVMuX74sN998s+k6c2wrV65063rzhGtZS5QoYW4dZduzZ4/pQnOl5XK1Y8cO053lWo7u3bub3C1H3Wl9vfnmm/L666/LQw89ZLocAeBaQq95BhAk9Au3fPnyzlwUzaXp2bOnjB49WgoVKiRr1qyRp59+2uTAuAYLrs6fPy/Nmzc3mwYwmjujAYreTyt3RoOjrVu3mlyZxYsXy7Bhw0y+kObnaF6PBlEaGE2fPt3krWiOSvXq1VM8n45Sc3AEcMn3ZSa/RQOinDlzmsBOb11p4JERqZU1I2XTsmhOT7t27VIc05wfh1WrVpmy6vup+T+hoXykAUgfnxKwhWXLlpkWjQEDBpj7+uWuX8Tjx4+XHDmuNHxq8rCrsLAw0wLiau/evaZVaOzYsSYhWm3evPmar69fyM2aNTPb8OHDTcCjZdLWon379pmgRxOXlQZgWUVbTi5evGgSotX69etNEOMou6vatWub69WWGUdZskPVqlVlw4YNbvu0XK40qVnrpWLFimk+z5w5c0wStQaUjz76qIwaNcoESwCQHgIfBB0dmXT8+HHzJX7ixAkzamrMmDGmhadTp07mHP1CvXTpkkyePNm0uKxdu1beffddt+fRLidteVi6dKmzy0i7tzQg0sfpCKldu3aZL9z0LFy4UH799Ve5++67pWDBgvLtt9+aoKty5crmvo7kmjZtmukS0tajwYMHZ1ldaKuRtmK98sorplVEg64+ffo4gz1X2sXVsWNHU0caEGogpKPE9Pq160pHTWWFvn37SqNGjUw3VZs2beT7778375ErbRXT90vrW0fAaXk1iNP6/s9//iO//fabaa3Tbi4dETdjxgxzvo5su+OOO7KknACC1NVcHyAoaEKv/lrrFhoaakYxNWvWzPrggw+cCb0OEyZMMMmymhjbvHlz66OPPjKPO336tPMcHSWlCc+6f/jw4Wafjs7SEUY62klHFn399dfmuI7ISs3q1atNEnHBggXNa9WsWdOaM2eO8/iSJUtM0rU+nx5bsWJFqgnDrs+/fPnyFGVNnoztSPQeNmyYuQZNau7evbsVFxeX5qiuhIQEc75eX65cuUz9tG3b1vrpp5/SrPNrlVXLqPu0zA46Mq506dKmPnQU25tvvpkikVxHsOkIMz1Hk8Pr169vTZs2zUpKSjLJz/qe6c8Ozz33nFWhQoU0k8wBQIXoP74OvgBkPZ3H559//sn0chIAEEwY1QUAAGyDwAcAANgGXV0AAMA2aPEBAAC2QeADAABsg8AHAADYBoEPAACwDQIfAABgGwQ+AADANgh8AACAbRD4AAAA2yDwAQAAYhf/H2x3RS1QlDD2AAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 600x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAj4AAAEiCAYAAAD5ztjRAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjAsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvlHJYcgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAOQtJREFUeJzt3QmczPX/wPH3suxiHesmR0JuchYqihISSSRFonLlKIXKlbSOXNFPPxQlcpSj/CJyk/uWM60jOSqsXGvXfv+P96dm/jO7s2s2a2d2vq/n4/E1M9/vd2Y+38+umfd+Pu/P5xNkWZYlAAAANpDO1wUAAABILQQ+AADANgh8AACAbRD4AAAA2yDwAQAAtkHgAwAAbIPABwAA2AaBDwAAsA0CHwAAYBsEPvAbQUFBydruvPNO87y6deuax0ePHhV/dPjwYenUqZOULFlSQkNDJSwsTO666y559NFH5b333pOff/45VcrhWmepKTo6WsaMGSM1a9aU7NmzS8aMGaVAgQJSrVo1efXVV+V///tfiryPXpteo79cNwD/FOzrAgAO7dq1S7Bv3bp1cuTIEalUqZLcc889bsdy584t/m7p0qXSrFkzuXr1qvmyf/jhhyVHjhzy66+/mmtbtmyZBAcHS9++fSUQXbhwQerVqyfbt2+XDBkymOCnYMGCcunSJdm2bZtMmDDB1EHjxo1TtVyrVq2Shx56yPzOTZs2LVXfG4BvEfjAb3j6AnrhhRdM4KPBw6BBgzw+7/PPP5crV67IHXfcIf5Eg53nn3/e3L711lum/Prl73D58mVZuHChZMmSJVXKs3//frf3Tw0DBgwwQY8GrYsWLUrwM9q8ebMsWbIk4K4bgP8i8EGaV6RIEfFH2qJz9uxZ82U/dOjQBMc14Hn22WdTrTylS5eW1Pb111+b25EjR3oMTGvUqGG2QLtuAP6LHB+keYnl+DhyO2JjY2XIkCFSokQJyZQpk5QpU0amTp3qPG/FihWm2yNbtmwSHh4ubdu2lT///NPje+lrTZw40XTZ6Pn6etqaMXbsWHPM1e+//25u8+TJk+xrsixLvvzyS9M1pmXS3CAtt7YaaetWUnUwc+ZMue+++yRr1qymWy1+fSTWKqKta4ULF5aQkBDJly+fPPPMM/LTTz95PP+7776TRx55xAQzer52X91///0yePDgW64D7YbSsmp5Tp06ZW61PFrXVapUMS18yRH/uvX19OetPvvsM7e8scRaFQEEDgIfBLyWLVvKqFGjpEKFCvLggw9KZGSkvPjiiyb4+eqrr6RBgwYmaNFbbYWZPn266VrT4MOVdllpQnKXLl3k0KFDJrjQL3/9cu7Vq5c89dRTEhcX5zxfgwi1Z88eWbt2rdfl1ddo06aNaQ3asmWLCawaNWpkusY0sNAvbS2LJxEREaZ7TROIH3/8cSlfvvxN32/BggVSuXJlEwRo3tQTTzwhxYoVkzlz5pjWmDVr1rid/9FHH5mcnJUrV5pgUq9b3+fYsWMJAgdHHXz88ccJ6vNmzp07Z+pYu8I0sHvggQdMXWpezq0EKBqg6c9aFS9e3LyeY4ufRwYgAFmAH2vXrp1+W1oDBw5M9Jw6deqYcyIjI9326z7dypcvb509e9a5f8WKFWZ/gQIFrFy5clmLFi1yHouKirLKlStnjut5rrp06WL2t2rVyrpw4YJz/8WLF61GjRqZYxMnTnTuj42NtcqWLWv2p0+f3mrcuLE1evRoa/Xq1dbly5cTvZ4RI0aY59StW9c6deqUc390dLTVoUMHc6xPnz4e6yA0NNRatWqVx9fV40WLFnXbp3WWJUsWKywszFq2bJnbscWLF1sZMmSwChcubN7boUiRIlZQUJC1ZcsWt/Pj4uKslStXuu2LiIhw/hxKly5t9e3b15o/f7514sSJRK9fX8PxnEceecS6dOmS89jmzZtNWdOlS2dt27bN7Xl6bZ4+0jxdt+M99PcLgL0Q+CDgA58ffvghwXMqV65sjj333HMJjo0bNy7Be545c8YZBFy5ciXBczRAyZgxo1WxYkW3/VqmWrVqOcvi2PS1mjRpkiB4iImJsXLnzm2CkdOnTyd4H33v/PnzW+Hh4daNGzcS1EHXrl0TrSdPAUCPHj3M/vHjx3t8Tvfu3c3xefPmOfdlypTJvL83tIxvvvmmud74daABpgaKrtfhGpRocHPgwIEEr6lBnx7XINAVgQ8Ab9DVhYCmo3m0myQ+nUdHaddVYse0C8s17yQmJkYee+wxk2sSX/78+c08PdoV49oNpbkl69evN1ufPn2kTp06Zh4ffa1vv/3W5Appl5KDjoD6448/pFatWiavJT5976pVq8r58+fN/EDxaTdVcofbq+bNm3s8rt1LjtFXDo7379ChQ6I5QA7p0qWT4cOHm9yjcePGmW4xRzK6Prdz584JuggdtNupVKlSCfa3bt3a3Can+xAAHAh8ENA0IEmfPn2C/Rp8KE8jjRzHdOI9B0fi9OTJkxOdUFG/yLWBQXNT4tNAZtiwYSaA0sRpHdpdtmxZk1v0yiuvmHltXN9H57ZJ7H0cE/5pgHSrI9wc76f14Om9nn766QTvpTk+mgP06aefmtwereNWrVrJ7Nmz5caNGx7fR5Ofu3fvbnKqNBdo3759Js/KkWOkidzxFS1a1ONrORKVf/vtt2RdKwAohrMjoGmLw60cd3C0SGgrhE6mmBQd5ZQUTTzW5GBtOdHkWp3k78cffzStT4730aTh2rVrJ/k6uXLlSrBPR38lh+P9PE0e6eree+913q9YsaIJXDTpWEd3aTCnrVa6aQuWPtZrTIqOUPvkk09My9H8+fNNMKcJ3QBwuxH4AF4oVKiQc0TQ+PHjU+Q1taVEAwCdwdjRouJ4H517JjVmFNb30wkiddSbp0AqMRpg6cg33ZS2dukotA0bNsiUKVPMyDdv6HB9DXw8tV5py5Anjv3aigQAyUVXF+AFHUKuXWbaRaX5Od642fBt7RbSofWuXW7Vq1c361mtXr3aY5dZStPh+EqDj1tRrlw56dq1q7m/d+9er+vAsU6Zpy7HnTt3esxjmjVrljMI/bccLVLx514CEPgIfAAv6Bez5qRoTowm1545c8bjl7hjpmKlycs6h5B2Y8Wnc/JoYq8GN9pyoV1Ejm6yN998U/766y+TcPzLL78keO7JkyfNXEMp4fXXXzcJ071795Z58+YlOK55TpqXo2uLKZ088cMPPzTdc/G7zBxLTzjm7nHkNul8SXq98WkQqfP7qBYtWiQ4rq+pi5i6TtjoWN9L84+0/v4tR2vRwYMH//VrAEib6OoCvKSjkjTw0eBGv+Q130eTifVLXXNeNPBp2rSpGaXk+OKeO3eu2bRbSycJ1FmYdTbjrVu3mvwWx4SJrjkxumDpgQMHzH7tCtPnaTLx9evXzRe1vpfm2ehEhbdKc4k0sVi7qbTc+ljfU8ulAZaOMtPr27Fjh+kW0zL06NHDBEqao6SJxrpPJ1o8ceKEefzyyy+7zQitAaN2femsy1pfOupNr0OvUenK9Z4WKdUJGHft2mXyoHTiyaioKDPLtra4vfPOO2Z1939Ly6l1qD8HnaRRW6y0RU9HxSV3ZByAtIXAB/CStowsXrxYZsyYYWY51q4YHeatyzHoCCQNRHSZBwcd+q5Ju99//73JfdHzNejR19FApn379makU/zRS5pwrcsyaCvIpEmTTFChAYgGTdqa8sYbb5hRVClFg7Xdu3fL6NGjzWgy3XQaAG0VadKkiWl50hFojhFvOqpr+fLlJijR52nQpgFNx44dpVu3bpIzZ07na+uszxok6vmaS6R1oF18OlRfr0+DooYNG3osl+Ycbdy40UwDoHV48eJFU46ePXuaZSdulQawWpc6LF5bkjRQ1eCOwAcIbEE6mY+vCwEADjoqTHOqdKRZaiR4A7AXcnwAAIBt0NUFAPBL2i3q7ShK2FuGDBk8TlbrCYEPAMCvaAbG6dOnE4weBJKSI0cOM5BER30mhRwfAIBf0XXyNOjJmzevZM6c+aZfZLA3y7LMtBdnz541wU+BAgWSPJ8WHwCAX3VvOYKe5MwmDnvL9M/i0Rr86O9OUt1eJDcDAPyGI6dHW3qA5HD8ztwsL4wWHy/o/B66EnTWrFlpcgWAFOia0NnJda6oxBYK5rMWyeXt7wyBjxc06HGdhh8AcOt0tm/Hwrxwn1lcJ+rULakveV1jz7FQcGo6ffq0mbBVl+PR0VTeJKEPGjRIFixYYCYxTYxOTKqvpefdTgQ+XtCWHsd/0mzZsvm6OACQpuks3PrHpOOz1RtxUeck7solSS3pModJuuz/Pwv5zST2pe2YkFOXqNHE27QecKkxY8aYBHQNYnRR5bSGwCcZzWca9BD4AEDqdk1o0BM1caDIjVhJNemDJXvnwckKfuziyJEjZq2+kiVLSlpEcjMAwK+Zlp7UDHrUjdjb1sKk68TpwrghISGmlWXUqFFJnn/48GGzUG9oaKhZr07X04tvz5498vDDD5vRTToaThcLvnTp/8tft27dBC05zZo1c657p8ePHTsmvXr1MgFpYkGpllfLr+sJ6jmO5x8/ftys+6fr+WkDQcuWLeXMmTNJjt577bXXTCuYlvfNN980uV+pgcAHAIBUogvialCgCxprsKK5L/379090XTodXKMLBetiwJs2bZKPP/7YLNzr6vLly9KgQQOzkLEuajx37lz54YcfzKLB3po3b57Jt3r33XdNN5Zunujr6wLMeg16zrhx40wZNeg5d+6crF692gRmv/zyS5KLKWuwp9f86aefyrp168xzNWcpNdDVBQBACli0aJFp8YjfsuFq9OjRUq9ePRPsqLvvvlv27dsnI0eOdLaeuNIA5sCBA/L999+bUXDq/fffl4YNGzrPmTlzply7ds20wmTJksXsmzBhgjRp0kSGDx8u+fLlu2nZc+bMaea+0bwrnf04MXny5DEtVdqy5DhPAx0N4iIjI50DgbQs2qqlgVL16tUTvM7YsWOlX79+JqhTGtDpNaYGWnwAAEgBmsSsCb+u25QpU9zO2b9/v9SuXdttnz7W7qz4QZLjfA0mHEGPqlmzZoJzKlWq5Ax6HK+pLTEHDx6U223/P2V0Hf2sXXLajaXH4ouKijKtRffee69zX3BwsFSrVu22l9W8V6q8CwAAAU4DjxIlSrjt+/XXX8Uf6HxJ8XNoYmy6ACwtPgAApJIyZcrI+vXr3fbpY+3y8rTMgp6vU6m45txs3LgxwTm7du0yuT6ur6nBTqlSpZxdVK6vcePGDdm7d6/b62gekadWJ2+uScuom4N23+nwfm35iU+HwOt6Wpqz5BAbG2vyn1IDgQ8AAKnk9ddfl+XLl8uQIUPk0KFD8tlnn5l8nN69e3s8v379+iYoateunQlu1q5dK2+//bbbOW3atDEjvvQcDWZWrlwpr776qplk0JHfoyO+/ve//5lNc4Y6d+6cYOJBHbG1Zs0aOXnypPzxxx9eX5OWsUKFCqYc27dvl82bN0vbtm2lTp06iXZf9ejRQ4YNG2bmPdLydOnSxauJEFMCgQ8AAKmkSpUqMmfOHJk1a5aUL19eBgwYYEZSeUpsVtpqo6Odrl69KjVq1JCOHTvK0KFDE6xRpYnBOjJKE4lbtGhhEqg1oHJ48cUXTWDU9p+A5K677jI5Sa60HEePHpXixYubFiJv6bD2hQsXmlFlOuxeAyF9/dmzZycZAGpgpmXSnCVNqn7yySclNQRZqTVwPo3PMqpNc5qQxQSGAHD7PlN1dJKODipWrJhpxVBMYAhvePrd8YTk5mToOea8ZAxNfv9nIOh9vJvIP/NZZYy7Zm4DaQnBoNBMzvuFRn/h07IAcKfBhwYh/rxkBdIOAh94JcS6JhLAbYPWtau+LgKAJGgQQiCClECODwAAsA3bBD4fffSRyVjXfj+dNEmzzgEAgL3YIvDRzHJdDG3gwIFmqJ3OcKnrmpw9e9bXRQMAAKnIFoGPro3y0ksvSfv27c1kSromiA7/08XRAACAfQR84HP9+nUzG6TOK+A6L4I+3rBhg8fnREdHm+GWrhsAAEj7Aj7w0dkndQru+KvT6uPTp097fE5ERISZY8KxuS68BgAA0q6AD3z+jX79+pmJtRyb6/ojAAAg7Qr4wCd37txm4bczZ8647dfH+fPn9/ickJAQM5uo6wYAQGrSkchjx471+vxVq1aZ5SNSY82rBQsWmJXo9fu1Z8+eKXY9Wn597dsp4Ccw1NVmq1atahaFa9asmdkXFxdnHnfr1s3XxQMAeOHMH9ESdTH1lqzIni1Y8uUO8epc/bJOio4oHjRoULLLsGXLFsmSJYvX59eqVcuswK4pGrfbK6+8YgYMde/e3ayzlZYEfOCjdCi7LoSmq8TqIm8acV6+fNn80AAA/h/0tO2xS67HpN708RkzBMnn4yp5FfxosOE6fYouPHrw4EHnvrCwMOd9XR5T806Dg2/+9ZuchUJNmTNmTLQnIyVdunTJTAej08IULFhQ0pqA7+pSrVq1kg8++MD8Mt5zzz2yc+dOWbJkSYKEZwCA/9GWntQMepS+n7ctTBpsODZtbdEWIMfjAwcOmBaRxYsXm94HTaVYt26dHDlyRJo2bWq+hzQw0lXVf/jhhyS7hvR1p0yZYlYx1ylZSpYsKd98802iXV3Tpk2THDlymJXby5QpY97nsccecwvUYmNjTauNnpcrVy7p06ePaShw9JDEp+/haOF5+OGHzfvpPvX1119LuXLlzDVq2UeNGpVkvR0+fNis5q4TC+tUM8uWLXP/GVy/bnpmChQoYM4pWrSoGXx0q2wR+CitvGPHjpmh6ps2bTKzN8N70UGhEp3u700/fgJt2S5dpNSxAUBK69u3rwwbNkz2798vFStWNK0mjRo1MmkXO3bsMAFJkyZN5Pjx40m+zuDBg6Vly5aye/du8/w2bdrIuXPnEj3/ypUr5g//6dOny5o1a8zr9+7d23l8+PDhMmPGDJk6daqsX7/eTN+SVI6Ndqc5WrM00NEgSvfptDFarmeeeUb27Nljuvb69+9vgi9PNOWkefPmppVKv5N1fj0Nulx9+OGHJrCbM2eOeU8tpwZUt8oWXV0pZWyvcBsnOs/wdQEAIM1699135ZFHHnE+zpkzp1lFwGHIkCEyf/5880WfVP7pCy+8IK1btzb333//fRMc6BJMGjh5EhMTY4KK4sWLm8fdunUzZXEYP368GcmsrUhqwoQJ8t133yX6/hqo5M2b13kNjq41nSi4Xr16JthRd999t+zbt09Gjhxpyhyftm5pa5i2Rjm6y/R6GjZs6DxHgzRt1br//vtNy5K2+KQE27T4AADgK5pj6kpbfLTlRbugtJtJu6G0NehmLT7aWuSgic/6x3hSyy9pl5gj6FEFChRwnq/TtegIZ819ddBRWtoll1xa9tq1a7vt08fanaU5TZ7O1znyXHOEatas6XaOBkyamlKqVCnTHbd06VJJCQQ+AADcZvFHZ2nQoy082sqxdu1a8wVfoUIFk9eSlAwZMrg91pYQ7TZKzvmWlTaSFapUqSKRkZGmNezq1aumK61Fixa3/LoEPgAApDLNp9EWDe1i0oBHu4yOHj2aqmXInj27Sa7WYfMO2jqji3knl7Zc6TW50sfa5aWtSJ7O18mBXROtN27cmOA8bdHSAUqTJ082I+Y0ryipnCZvkOMDAEAq09yVefPmmYRmbYXR3JikWm5ul1dffdWMlNLJCEuXLm1yfs6fP3/TuYnie/31183ING2d0UBF18LUfKH//Oc/Hs/X9TI1KNIRZJoHpEnVb7/9tts5mjekXXOVK1c2a2zOnTvXBIjaNXgraPEBACCV6Zd6eHi4GRGlwY/OiaNdO6mtT58+Jlm6bdu2JsdGc420LDp8PDm07Dr6atasWVK+fHkzfYwmUXtKbFYayGhXn3ZhaY5Rx44dZejQoW7n6LD5ESNGmPwoDaq0RUwTr/W5tyLISiudfT6kkag2CWoimH1HdQHA7f9MvXbtmsnrKFasmPPL198nMAwkcXFxphtK82m09SYt8fS74wldXQAAv6bBhwYh/rpkRVp27NgxM1qqTp06Zp477Z7S4OHZZ5+VQEXgAwDwexqE2CEQSW3p0qUzkwzqKDPtANJuKp1jR1t9AhWBDwAANlW4cOEEo7ECHcnNAADANgh8AACAbRD4AAD8DgOOcbt+Zwh8AAB+w7HEgq4qDiSH43cm/jId8ZHcnAw9x5yXjKEJF1tD2tX7+D+rIP8zSWnGuGuOuwEhKDST2+NCo7/wWVkAb+jyBjozr2MhTV1kM7mzCMN+LT1XrlwxvzP6u+NpiQxXBD6wtRDr2t93ArRV3bp21ddFAJJNlyVQSa06DsSnQY/jd0fsHvisWbPGrAWybds2syCaTpPdrFkzXxcLAOCBtvDoGk158+aVmJgYXxcHaYB2b92spcdWgc/ly5elUqVK8uKLL0rz5s19XRwAgBf0i8zbLzPAW7YIfBo2bGg2AABgb4zqAgAAtmGLFp/k0oXadHNdSRgAAKR9tPh4EBERIdmzZ3duupYJAABI+wh8POjXr59ERUU5txMnTvi6SAAAIAXQ1eVBSEiI2QAAQGCxReBz6dIl+fnnn52PIyMjZefOnZIzZ04pUqSIT8sGAABSjy0Cn61bt8pDDz3kfPzaa6+Z23bt2sm0adN8WDIAAJCabBH41K1bl5V+AQCAPQIfIDHRQaG2WqQUAOwuyKIp5KZ0Hh8d1q4jvLJly+br4gBAmsZnKnyJ4ewAAMA2CHwAAIBtEPgAAADbIPABAAC2QeADAABsg8AHAADYBoEPAACwjX81geHhw4dl5cqVcvbsWYmLi3M7NmDAgJQqGwAAgG8Dn8mTJ0vnzp0ld+7ckj9/fgkK+v95bvU+gQ8AAAiYwOe9996ToUOHSp8+fW5PiQAAAPwlx+f8+fPy9NNP357SAAAA+FPgo0HP0qVLb09pAAAA/Kmrq0SJEtK/f3/ZuHGjVKhQQTJkyOB2vHv37hKoeo45LxlDb/i6GIBXeh/v5rbqvArklecLjf7CZ2UBEMCBz6RJkyQsLExWr15tNlea3BzIgQ+QloRY10QsCVjWtau+LgIAOwQ+kZGRt6ckAAAA/jyBoWVZZvNnERERUr16dcmaNavkzZtXmjVrJgcPHvR1sQAAQFoJfD7//HOT35MpUyazVaxYUaZPny7+SLvjunbtanKSli1bJjExMfLoo4/K5cuXfV00AADg711do0ePNsnN3bp1k9q1a5t969atk06dOskff/whvXr1En+yZMkSt8fTpk0zLT/btm2TBx980GflAgAAaSDwGT9+vEycOFHatm3r3PfEE09IuXLlZNCgQX4X+MQXFRVlbnPmzJnoOdHR0WZzuHjxYqqUDQAA+FlX16lTp6RWrVoJ9us+PebPdF2xnj17mpaq8uXLJ5kXlD17dudWuHDhVC0nAADwk8BH5/GZM2dOgv2zZ8+WkiVLij/TXJ+9e/fKrFmzkjyvX79+pmXIsZ04cSLVyggAAPyoq2vw4MHSqlUrWbNmjTPHZ/369bJ8+XKPAZG/0JykRYsWmXIXKlQoyXNDQkLMBgAAbN7i89RTT8mmTZvM6uwLFiwwm97fvHmzPPnkk+JvdLi9Bj3z58+XFStWSLFixXxdJAAAkFZafFTVqlXliy/SxvTw2r01c+ZMWbhwoZnL5/Tp02a/5u7oUHwAAGAfXgU+OqopW7ZszvtJcZznL3QEmqpbt67b/qlTp8oLL7zgo1IBAAC/DXzCw8PNiC2d/yZHjhxmTS5PXUq6/8YN/1rE099nlgZul+igUFstUgoAKRb4aG6MY96blStXil2N7RXudy1aQOJm+LoAAJA2A586deo472tysM5rE7/VR1tWGPYNAAACalSXBj6///57gv3nzp1jxBQAAAiswMeRyxPfpUuXJDQ0NKXKBQAA4Lvh7K+99pq51aBHFynNnDmz85gmNOvcPvfcc0/KlxAAACC1A58dO3Y4W3z27NkjGTNmdB7T+5UqVZLevXunVLkAAAB8F/g4RnO1b99exo0bx+gmAAAQ+DM368R/AAAAtlmyYuvWrWZB0uPHj8v169fdjs2bNy+lygYAAODbUV2zZs2SWrVqyf79+83CnzExMfLTTz+ZSQ51/SsAAICACXzef/99GTNmjHz77bcmqVnzfQ4cOCAtW7aUIkWK3J5SAgAA+CLwOXLkiDRu3Njc18Dn8uXLZoh7r169ZNKkSSlRJgAAAP8IfHTB0r/++svcv+OOO2Tv3r3m/oULF+TKlSspX0IAAABfJTc/+OCDsmzZMqlQoYI8/fTT0qNHD5Pfo/vq1auXUuUCAABIcUGWzkiYDLom17Vr16RgwYISFxcnI0aMkB9//FFKliwp77zzjmkRCjQXL140idvtBx2VjKHMXwT4Wu/j3UT+WTknY9w1c5twIZ20Kyg0k/N+odFfSKB+pkZFRTEnHPy7xSc2NlYWLVokDRo0MI/TpUsnffv2vV1lAwCPQqxrIsn6ky1tsa5d9XURgICVrByf4OBg6dSpk2nxSSsmTpwoFStWNH9V6FazZk1ZvHixr4sFAADSQnJzjRo1ZOfOnZJWFCpUSIYNGybbtm0zEy8+/PDD0rRpUzP3EAAAsJdkJzd36dLFrNR+4sQJqVq1qmTJksXtuLau+JMmTZq4PR46dKhpBdq4caOUK1fOZ+UCAABpIPB55plnzG337t2d+3QeH82R1tsbN26Iv9KyzZ0718w9pF1eAADAXpId+ERGRkpas2fPHhPoaG5SWFiYWWqjbNmyiZ4fHR1tNtcRCAAAwIaBT9GiRSWtKVWqlMlL0qGTX331lbRr105Wr16daPATEREhgwcPTvVyAgAAP0tuVtOnT5fatWubuXyOHTtm9o0dO1YWLlwo/kiX1ihRooTJSdKgplKlSmaNscT069fPBEmOTfOZAACADQMfTQzW5OZGjRqZZSocOT05cuQwwU9aoBMvunZlxRcSEuIc/u7YAACADQOf8ePHy+TJk+Xtt9+W9OnTO/dXq1bN5NL4G229WbNmjRw9etSUTx+vWrVK2rRp4+uiAQCAtJDcXLlyZY+tJDpayt+cPXtW2rZtK6dOnTJTpOtw+++//14eeeQRXxcNAAD4e+BTrFgxkygcP8l5yZIlUqZMGfE3n3zyia+LAAAA0mrgo/k9Xbt2NUPDde6ezZs3y5dffmmShqdMmXJ7SgkALqKDQm2zSCkAH6/OrmbMmCGDBg2SI0eOmMc6ukuHf3fo0EECESsJA0DK4TMVaS7wcbhy5YpcunRJ8ubNK4GM/6QAkHL4TEWaGtWli3zqMHaVOXNmZ9Cjv8h6DAAAIGACHx0Kfv369QT7Nedn7dq1KVUuAAAA3yU3796923l/3759cvr0aedjncRQR3XdcccdKV9CAACA1A587rnnHrP6um6eurQyZcpkJjcEAABI84GPTlyoedB33XWXGcKeJ08et7WwNNfHdSZnAACANBv4OCYs1HWuAAAAAjrw+eabb7w674knnriV8gAAAPg+8GnWrNlNz9H8H8dq7QAAAGk28KGLCwAA2G4eHwAAgLSKwAcAANhGsldnt7OeY85LxlBymADcHr2Pd/v7jsvK84G46vxf0TG+LgpsjMAHAPxEiHXt7zv/eulo/2Zdu/r37XUCH6SBrq5AGa01bNgwM/qsZ8+evi4KAADw18BH1+Hq27evHDp0SNKqLVu2yH//+1+pWLGir4sCAAD8OfDp2rWrfPXVV1KmTBl54IEHZNq0aXLlyhVJKy5duiRt2rSRyZMnS3h4uK+LAwAA/Dnw6d+/v/z888+yfPlys15Xt27dpECBAvLSSy/Jpk2bxN9p4Na4cWOpX7++r4sCAADSynD2unXrymeffSanT5+WUaNGyf79+6VmzZpSrlw5GT16tPijWbNmyfbt2yUiIsKr86Ojo+XixYtuGwAAsPE8PmFhYdKxY0dZt26dfPvttyYQeuONN8TfnDhxQnr06CEzZsyQ0NBQr56jAVL27NmdW+HChW97OQEAgB8HPprfo3k+derUMQuT5sqVS4YOHSr+Ztu2bXL27FmpUqWKBAcHm2316tXy4YcfmvueRqv169dPoqKinJsGTwAAwIbz+Pz444/y6aefyty5cyU2NlZatGghQ4YMkQcffFD8Ub169WTPnj1u+9q3by+lS5eWPn36SPr06RM8JyQkxGwAAMCmgc+IESNk6tSpZjh7tWrVZOTIkdK6dWvJmjWr+DMtX/ny5d32ZcmSxbRQxd8PAAACm9eBjwY6zz33nGnpIWAAAAABHfj89ttvkiFDBgkEq1at8nURAACAPyc3r127VsqWLetxaLcmAOtwdj0HAPDvRAeF/r2l+3sLtCW7dJFSs4X8vVgp4AtBlmV59X9LR2499NBD0qtXL4/HdZTUypUrZf78+RJoNNjTYe0a4GXLls3XxQGANI3PVKSJFp9du3bJY489lujxRx991AwdBwAASPOBz5kzZ5LM8dE5cX7//feUKhcAAIBvV2ffu3dvosd3795t1u4CAABI84FPo0aNzEKl165dS3Ds6tWrMnDgQHn88cdTunwAAACpn9ysXV267IPOdKwrs5cqVcrsP3DggHz00Udm6QddCDRfvnwSaEjEA4CUw2cq0sQ8PhrQ6HIVnTt3NmtZOeKloKAgadCggQl+AjHoAQAANl2rq2jRovLdd9/J+fPn5eeffzbBT8mSJSU8PPz2lRAAAMBXi5QqDXSqV6+eUmUAAADwr+RmAACAtI7ABwAA2AaBDwAAsA0CHwAAYBv/KrnZrubuOimZwxKuTn87NY/60Hk/WK6b26BULQE8WTL1h3/uBUlsTKyPSwOkDcGZspjbK9dv+LoosDECHz+X4Z9gB/4lNoYPbiC5Yq9e/vuW/z/wIb/s6jp69KiZGHHnzp2+LgoAAAggfhn4AAAA3A4EPgAAwDZ8GvjExcXJiBEjpESJEhISEiJFihSRoUOHJjhPF0Dt0KGDFCtWTDJlymQWSB03bpzbOatWrZIaNWpIlixZJEeOHFK7dm05duyYObZr1y556KGHJGvWrGZBvKpVq8rWrVtT7ToBAIB/8Glysy52OnnyZBkzZozcf//9curUKbPau6cAqVChQjJ37lzJlSuXWSz15ZdflgIFCkjLli0lNjZWmjVrJi+99JJ8+eWXcv36ddm8ebPJE1Jt2rSRypUry8SJE83q8po7lCFDhkTLFR0dbTbXlYQBAEDa57PA56+//jKtNhMmTJB27dqZfcWLFzcBkCY3u9IgZfDgwc7H2vKzYcMGmTNnjgl8NDCJioqSxx9/3LyGKlOmjPP848ePyxtvvCGlS5c2j3Vh1aRERES4vR8AAAgMPuvq2r9/v2lVqVevnlfnf/TRR6aLKk+ePBIWFiaTJk0yAY3KmTOnvPDCC9KgQQNp0qSJCai09cjhtddek44dO0r9+vVl2LBhcuTIkZu2RGkg5dhOnDhxi1cLAABsHfhoro63Zs2aJb179zZ5PkuXLjVdVe3btzddWg5Tp041rUC1atWS2bNny9133y0bN240xwYNGiQ//fSTNG7cWFasWCFly5aV+fPnJ/p+mm+kuUCuGwAASPt8Fvhod5MGP8uXL7/puevXrzcBTZcuXUyujiZDe2q10WPaWqM5QOXLl5eZM2c6j2kg1KtXLxM4NW/e3ARKAADAXnwW+ISGhkqfPn3kzTfflM8//9wEMtpC88knn3gMknQU1vfffy+HDh2S/v37y5YtW5zHIyMjTcCjLT46kkuDm8OHD5s8n6tXr0q3bt3MqC89pkGUPtc1BwgAANiDT0d1aQATHBwsAwYMkN9++82M0urUqVOC81555RXZsWOHtGrVyozUat26tWn9Wbx4sTmeOXNmMxrss88+kz///NO8TteuXc3zdMSX7mvbtq2cOXNGcufObVp8SF4GAMB+gizLsnxdCH+no8ayZ88uU9bsk8xhWVP1vVmk1D+xSClwa4uUPv3FTjN4hBxKpDYCn2QEPvwnBYBbx2cqfIklKwAAgG0Q+AAAANsg8AEAALZB4AMAAGyDwAcAANgGgQ8AALANAh8AAGAbBD4AAMA2CHwAAIBtEPgAAADbIPABAAC2QeADAABsg8AHAADYBoEPAACwDQIfAABgGwQ+AADANgh8AACAbRD4AAAA2yDwAQAAtkHgAwAAbIPABwAA2AaBDwAAsA0CHwAAYBsEPgAAwDYIfAAAgG0Q+AAAANsg8AEAALZB4AMAAGyDwAcAANgGgQ8AALANAh8AAGAbBD4AAMA2CHwAAIBtEPgAAADbIPABAAC2QeADAABsg8AHAADYBoEPAACwjWBfFyAtsCzL3F68eNHXRQGANM/xWer4bAVSE4GPF/78809zW7hwYV8XBQAC6rM1e/bsvi4GbIbAxws5c+Y0t8ePH+c/6T9/rWkQeOLECcmWLZvYHfXhjvpIiDpxFxUVJUWKFHF+tgKpicDHC+nS/Z0KpUEPH1r/T+uC+vh/1Ic76iMh6sTzZyuQmvitAwAAtkHgAwAAbIPAxwshISEycOBAcwvqIz7qwx31kRB14o76gC8FWYwnBAAANkGLDwAAsA0CHwAAYBsEPgAAwDYIfG7io48+kjvvvFNCQ0Pl3nvvlc2bN4sdRERESPXq1SVr1qySN29eadasmRw8eNDtnGvXrknXrl0lV65cEhYWJk899ZScOXNG7GDYsGESFBQkPXv2tHV9nDx5Up577jlzzZkyZZIKFSrI1q1bncc1hXDAgAFSoEABc7x+/fpy+PBhCUQ3btyQ/v37S7Fixcy1Fi9eXIYMGeK2LEMg18eaNWukSZMmUrBgQfN/Y8GCBW7Hvbn2c+fOSZs2bcxcRzly5JAOHTrIpUuXUvlKEOgIfJIwe/Zsee2118zog+3bt0ulSpWkQYMGcvbsWQl0q1evNl/iGzdulGXLlklMTIw8+uijcvnyZec5vXr1km+//Vbmzp1rzv/tt9+kefPmEui2bNki//3vf6VixYpu++1WH+fPn5fatWtLhgwZZPHixbJv3z4ZNWqUhIeHO88ZMWKEfPjhh/Lxxx/Lpk2bJEuWLOb/kAaJgWb48OEyceJEmTBhguzfv9881usfP368LepDPxv0M1L/WPTEm2vXoOenn34ynzmLFi0ywdTLL7+cilcBW9BRXfCsRo0aVteuXZ2Pb9y4YRUsWNCKiIiw7Obs2bP6Z6u1evVq8/jChQtWhgwZrLlz5zrP2b9/vzlnw4YNVqD666+/rJIlS1rLli2z6tSpY/Xo0cO29dGnTx/r/vvvT/R4XFyclT9/fmvkyJHOfVpPISEh1pdffmkFmsaNG1svvvii277mzZtbbdq0sV196O/9/PnznY+9ufZ9+/aZ523ZssV5zuLFi62goCDr5MmTqXwFCGS0+CTi+vXrsm3bNtMc6zq9uj7esGGD2HFtHeVYW0frRluBXOundOnSZv2dQK4fbQVr3Lix23XbtT6++eYbqVatmjz99NOmO7Ry5coyefJk5/HIyEg5ffq0W53osi/aZRyIdVKrVi1Zvny5HDp0yDzetWuXrFu3Tho2bGjL+nDlzbXrrXZv6e+Ug56vn7vaQgSkFNbqSsQff/xh+uzz5cvntl8fHzhwQOwkLi7O5LJot0b58uXNPv0Qy5gxo/mgil8/eiwQzZo1y3R5aldXfHasj19++cV07Wh38FtvvWXqpXv37qYe2rVr57xuT/+HArFO+vbtaxYj1YA3ffr05vNj6NChpvtG2a0+XHlz7XqrAbSr4OBg88dWoNcPUheBD7xq5di7d6/569WudFXtHj16mNwDTXTH3wGx/nX+/vvvm8fa4qO/J5rDoYGP3cyZM0dmzJghM2fOlHLlysnOnTvNHwya7GvH+gD8FV1dicidO7f5qy3+qBx9nD9/frGLbt26mSTDlStXSqFChZz7tQ60O/DChQu2qB/tytKk9ipVqpi/QnXTBGZN1tT7+pernepD6eicsmXLuu0rU6aMHD9+3Nx3XLdd/g+98cYbptXnmWeeMaPbnn/+eZPwriMk7Vgfrry5dr2NP3AkNjbWjPQK9PpB6iLwSYQ211etWtX02bv+hauPa9asKYFO8xM16Jk/f76sWLHCDNF1pXWjo3lc60eHu+uXXiDWT7169WTPnj3mr3jHpq0d2o3huG+n+lDa9Rl/igPNbylatKi5r78z+oXlWifaFaT5GoFYJ1euXDH5KK70jyf93LBjfbjy5tr1Vv9w0D8yHPSzR+tPc4GAFOPr7Gp/NmvWLDPqYNq0aWbEwcsvv2zlyJHDOn36tBXoOnfubGXPnt1atWqVderUKed25coV5zmdOnWyihQpYq1YscLaunWrVbNmTbPZheuoLjvWx+bNm63g4GBr6NCh1uHDh60ZM2ZYmTNntr744gvnOcOGDTP/ZxYuXGjt3r3batq0qVWsWDHr6tWrVqBp166ddccdd1iLFi2yIiMjrXnz5lm5c+e23nzzTVvUh4543LFjh9n0q2X06NHm/rFjx7y+9scee8yqXLmytWnTJmvdunVmBGXr1q19eFUIRAQ+NzF+/HjzZZYxY0YzvH3jxo2WHegHl6dt6tSpznP0A6tLly5WeHi4+cJ78sknTXBk18DHjvXx7bffWuXLlzd/IJQuXdqaNGmS23Edxty/f38rX7585px69epZBw8etALRxYsXze+Dfl6EhoZad911l/X2229b0dHRtqiPlStXevzM0IDQ22v/888/TaATFhZmZcuWzWrfvr0JqICUxOrsAADANsjxAQAAtkHgAwAAbIPABwAA2AaBDwAAsA0CHwAAYBsEPgAAwDYIfAAAgG0Q+AAAANsg8AFsrG7dumYF8VsRFBQkCxYsuKXXmDZtmuTIkeOWXgMAvEHgg4DywgsvmC9i3XTRUF01/ZFHHpFPP/3UuVikt/gy9s6pU6ekYcOGvi4GAHiFwAcB57HHHjNfxkePHpXFixfLQw89JD169JDHH39cYmNjfV28gKOrboeEhPi6GADgFQIfBBz9EtYv4zvuuEOqVKkib731lixcuNAEQdqK4zB69GipUKGCZMmSRQoXLixdunSRS5cumWOrVq2S9u3bS1RUlLMFadCgQebY9OnTpVq1apI1a1bzPs8++6ycPXs2yTL95z//kZIlS0poaKhphWrRooXz2JIlS+T+++83rUu5cuUyAdqRI0ecxzWA0/efM2eOPPDAA5IpUyapXr26HDp0SLZs2WLKEhYWZlpdfv/9d7fWr2bNmsngwYMlT548ki1bNunUqZNcv3490XJGR0dL7969Td1pvdx7772mLrzt6nKUdd68eSbgzJw5s1SqVEk2bNjg9hz9ORQpUsQcf/LJJ+XPP/9M8Lr6M9Ofn9bZXXfdZa7DEbi+++67UrBgQbfnNW7c2Lxnclv2ANhMii55CviYrgTdtGlTj8cqVapkNWzY0Pl4zJgx1ooVK6zIyEhr+fLlVqlSpazOnTubY7qi9tixY80K0brCum6OVaI/+eQT67vvvrOOHDlibdiwwapZs6bb68a3ZcsWK3369NbMmTOto0ePWtu3b7fGjRvnPP7VV19ZX3/9tXX48GFrx44dVpMmTawKFSpYN27cMMe1fPpfVVc/X7JkibVv3z7rvvvus6pWrWrVrVvXWrdunXnNEiVKWJ06dXKrC13lulWrVtbevXutRYsWWXny5LHeeuutRFeY79ixo1WrVi1rzZo11s8//2yNHDnSrKR96NChRK9PyzZ//vwEZdX309W3W7RoYRUtWtSKiYkx52zcuNFKly6dNXz4cHNc6yJHjhxW9uzZna+p7691P23aNFPPS5cute68805r0KBB5nhsbKyp92bNmpnHEyZMMK9x7NixRMsJAIrAB7YJfDQAKFOmTKLPnTt3rpUrVy7n46lTp7p9GScV2OiXvSMwik+DGv0Sv3jxolfX8Pvvv5vX27Nnj1swMWXKFOc5X375pdmnAZtDRESECd5c6yJnzpzW5cuXnfsmTpxogiFHUOUa+GjQoAHayZMn3cpTr149q1+/fskKfFzL+tNPP5l9+/fvN49bt25tNWrUKMHPxrWu9T3ff/99t3OmT59uFShQwPlYA6KsWbNaffr0sTJlymTNmDEjiVoFgL/R1QXb0O9o7YZx+OGHH6RevXqmW0e7rZ5//nnTdXLlypUkX2fbtm3SpEkT01Wjz6tTp47Zf/z4cY/na3J10aJFTXeNvseMGTPc3uPw4cPSunVrc1y7o+68806Pr1exYkXnfe0uU9pV57ovfpebdjNpd5JDzZo1TXfeiRMnEpRzz549cuPGDbn77rtN15ljW716tVvXmzdcy1qgQAFz6yjb/v37TReaKy2Xq127dpnuLNdyvPTSSyZ3y1F3Wl8ffPCBDB8+XJ544gnT5QgANxN80zOAAKFfuMWKFXPmomguTefOnWXo0KGSM2dOWbdunXTo0MHkwLgGC64uX74sDRo0MJsGMJo7owGKPk4sd0aDo+3bt5tcmaVLl8qAAQNMvpDm52hejwZRGhhNnjzZ5K1ojkr58uUTvJ6OUnNwBHDx991KfosGROnTpzeBnd660sAjOTyVNTll07JoTk/z5s0THNOcH4c1a9aYsurPU/N/goP5SAOQND4lYAsrVqwwLRq9evUyj/XLXb+IR40aJenS/d3wqcnDrjJmzGhaQFwdOHDAtAoNGzbMJESrrVu33vT99Qu5fv36Zhs4cKAJeLRM2lp08OBBE/Ro4rLSACylaMvJ1atXTUK02rhxowliHGV3VblyZXO92jLjKMvtUKZMGdm0aZPbPi2XK01q1nopUaJEoq8ze/Zsk0StAWXLli1lyJAhJlgCgKQQ+CDg6Mik06dPmy/xM2fOmFFTERERpoWnbdu25hz9Qo2JiZHx48ebFpf169fLxx9/7PY62uWkLQ/Lly93dhlp95YGRPo8HSG1d+9e84WblEWLFskvv/wiDz74oISHh8t3331ngq5SpUqZxzqSa9KkSaZLSFuP+vbtm2J1oa1G2or1zjvvmFYRDbq6devmDPZcaRdXmzZtTB1pQKiBkI4S0+vXrisdNZUSunfvLrVr1zbdVE2bNpXvv//e/IxcaauY/ry0vnUEnJZXgzit7/fee09+/fVX01qn3Vw6Im7q1KnmfB3Zdt9996VIOQEEqH9yfYCAoAm9+mutW3BwsBnFVL9+fevTTz91JvQ6jB492iTLamJsgwYNrM8//9w87/z5885zdJSUJjzr/oEDB5p9OjpLRxjpaCcdWfTNN9+Y4zoiy5O1a9eaJOLw8HDzXhUrVrRmz57tPL5s2TKTdK2vp8dWrVrlMWHY9fVXrlyZoKzxk7Edid4DBgww16BJzS+99JJ17dq1REd1Xb9+3Zyv15chQwZTP08++aS1e/fuROv8ZmXVMuo+LbODjowrVKiQqQ8dxfbBBx8kSCTXEWw6wkzP0eTwGjVqWJMmTbLi4uJM8rP+zPS+w6uvvmoVL1480SRzAFBB+o+vgy8AKU/n8blw4cItLycBAIGEUV0AAMA2CHwAAIBt0NUFAABsgxYfAABgGwQ+AADANgh8AACAbRD4AAAA2yDwAQAAtkHgAwAAbIPABwAA2AaBDwAAsA0CHwAAIHbxf0y9uQgRoi+UAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 600x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for cv in cvs:\n",
    "    this_cv = cv(n_splits=n_splits)\n",
    "    fig, ax = plt.subplots(figsize=(6, 3))\n",
    "    plot_cv_indices(this_cv, X, y, ax, n_splits)\n",
    "\n",
    "    ax.legend(\n",
    "        [Patch(color=cmap_cv(0.8)), Patch(color=cmap_cv(0.02))],\n",
    "        [\"Holdout fold\", \"Training folds\" ],\n",
    "        loc=(1.02, 0.8),\n",
    "    )\n",
    "    # Make the legend fit\n",
    "    plt.tight_layout()\n",
    "    fig.subplots_adjust(right=0.7)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ml2025",
   "language": "python",
   "name": "ml2025"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}