anon-ae.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandas import read_csv, set_option, get_dummies, DataFrame\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import cross_validate\n",
    "from sklearn.inspection import permutation_importance\n",
    "from numpy import mean, max, prod, array, hstack\n",
    "from numpy.random import choice\n",
    "from matplotlib.pyplot import barh, yticks, ylabel, xlabel, title, show, scatter, cm, figure, imshow\n",
    "from tensorflow.keras.layers import Input, Dense, Dropout, Activation, BatchNormalization\n",
    "from tensorflow.keras import Model\n",
    "from tensorflow.keras.datasets import mnist\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from tensorflow.keras.utils import plot_model\n",
    "from tqdm import tqdm\n",
    "\n",
    "\n",
    "set_option('display.max_columns', None)\n",
    "set_option('display.max_rows', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_autoencoder(dim_input, dim_layer_1, dim_layer_2):\n",
    "    \n",
    "    input_layer = Input(shape=(dim_input,))\n",
    "    x = Activation(\"relu\")(input_layer)\n",
    "    x = Dense(dim_layer_1)(x)\n",
    "    x = Activation(\"relu\")(x)\n",
    "    bottleneck_layer = Dense(dim_layer_2)(x)\n",
    "    x = Activation(\"relu\")(bottleneck_layer)\n",
    "    x = Dense(dim_layer_1)(x)\n",
    "    x = Activation(\"relu\")(x)    \n",
    "    output_layer = Dense(dim_input, activation='relu')(x)\n",
    "    \n",
    "    encoder = Model(input_layer, bottleneck_layer)\n",
    "    autoencoder = Model(input_layer, output_layer)\n",
    "    autoencoder.compile(optimizer='adam', loss='mse')\n",
    "    \n",
    "    return autoencoder, encoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = X_train.astype('float32') / 255.\n",
    "X_test = X_test.astype('float32') / 255.\n",
    "X_train = X_train.reshape((len(X_train), prod(X_train.shape[1:])))\n",
    "X_test = X_test.reshape((len(X_test), prod(X_test.shape[1:])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "autoencoder, encoder = build_autoencoder(\n",
    "    dim_input=X_train.shape[1], \n",
    "    dim_layer_1=64,\n",
    "    dim_layer_2=16\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 784)]             0         \n",
      "_________________________________________________________________\n",
      "activation (Activation)      (None, 784)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 64)                50240     \n",
      "_________________________________________________________________\n",
      "activation_1 (Activation)    (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 16)                1040      \n",
      "_________________________________________________________________\n",
      "activation_2 (Activation)    (None, 16)                0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 64)                1088      \n",
      "_________________________________________________________________\n",
      "activation_3 (Activation)    (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 784)               50960     \n",
      "=================================================================\n",
      "Total params: 103,328\n",
      "Trainable params: 103,328\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "autoencoder.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_model(autoencoder)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "callbacks = [\n",
    "    EarlyStopping(\n",
    "        monitor='val_loss', \n",
    "        min_delta=0.0001,\n",
    "        patience=5, \n",
    "        restore_best_weights=True\n",
    "    )\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0520 - val_loss: 0.0351\n",
      "Epoch 2/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0326 - val_loss: 0.0311\n",
      "Epoch 3/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0298 - val_loss: 0.0294\n",
      "Epoch 4/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0283 - val_loss: 0.0281\n",
      "Epoch 5/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0275 - val_loss: 0.0278\n",
      "Epoch 6/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0270 - val_loss: 0.0270\n",
      "Epoch 7/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0265 - val_loss: 0.0266\n",
      "Epoch 8/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0257 - val_loss: 0.0255\n",
      "Epoch 9/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0249 - val_loss: 0.0250\n",
      "Epoch 10/100\n",
      "165/165 [==============================] - 1s 6ms/step - loss: 0.0245 - val_loss: 0.0247\n",
      "Epoch 11/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0242 - val_loss: 0.0243\n",
      "Epoch 12/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0239 - val_loss: 0.0241\n",
      "Epoch 13/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0237 - val_loss: 0.0239\n",
      "Epoch 14/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0235 - val_loss: 0.0235\n",
      "Epoch 15/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0228 - val_loss: 0.0230\n",
      "Epoch 16/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0225 - val_loss: 0.0226\n",
      "Epoch 17/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0223 - val_loss: 0.0224\n",
      "Epoch 18/100\n",
      "165/165 [==============================] - 1s 5ms/step - loss: 0.0221 - val_loss: 0.0223\n",
      "Epoch 19/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0219 - val_loss: 0.0223\n",
      "Epoch 20/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0218 - val_loss: 0.0220\n",
      "Epoch 21/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0216 - val_loss: 0.0220\n",
      "Epoch 22/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0215 - val_loss: 0.0218\n",
      "Epoch 23/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0214 - val_loss: 0.0216\n",
      "Epoch 24/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0213 - val_loss: 0.0215\n",
      "Epoch 25/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0212 - val_loss: 0.0215\n",
      "Epoch 26/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0211 - val_loss: 0.0213\n",
      "Epoch 27/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0210 - val_loss: 0.0213\n",
      "Epoch 28/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0209 - val_loss: 0.0213\n",
      "Epoch 29/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0209 - val_loss: 0.0212\n",
      "Epoch 30/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0208 - val_loss: 0.0212\n",
      "Epoch 31/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0207 - val_loss: 0.0211\n",
      "Epoch 32/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0207 - val_loss: 0.0210\n",
      "Epoch 33/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0206 - val_loss: 0.0210\n",
      "Epoch 34/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0206 - val_loss: 0.0209\n",
      "Epoch 35/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0205 - val_loss: 0.0208\n",
      "Epoch 36/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0205 - val_loss: 0.0209\n",
      "Epoch 37/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0204 - val_loss: 0.0208\n",
      "Epoch 38/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0204 - val_loss: 0.0209\n",
      "Epoch 39/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0203 - val_loss: 0.0207\n",
      "Epoch 40/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0203 - val_loss: 0.0206\n",
      "Epoch 41/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0203 - val_loss: 0.0207\n",
      "Epoch 42/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0202 - val_loss: 0.0206\n",
      "Epoch 43/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0202 - val_loss: 0.0206\n",
      "Epoch 44/100\n",
      "165/165 [==============================] - 1s 4ms/step - loss: 0.0202 - val_loss: 0.0207\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7f169de600d0>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "autoencoder.fit(\n",
    "    X_train, X_train,\n",
    "    epochs=100,\n",
    "    batch_size=256,\n",
    "    shuffle=True,\n",
    "    validation_split=0.3,\n",
    "    callbacks=callbacks\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoded = array(encoder(X_test))\n",
    "decoded = array(autoencoder(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x576 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = figure(figsize=(16, 8))\n",
    "\n",
    "n_plots = 10\n",
    "n_rows = int(n_plots/2)\n",
    "\n",
    "for j in range(n_plots):\n",
    "    \n",
    "    fig.add_subplot(n_rows, 6, 3*j+1)\n",
    "    plot_tmp = imshow(X_test[j].reshape([28, 28]))\n",
    "    plot_tmp.axes.get_xaxis().set_visible(False)\n",
    "    plot_tmp.axes.get_yaxis().set_visible(False)\n",
    "    \n",
    "    fig.add_subplot(n_rows, 6, 3*j+2)\n",
    "    plot_tmp = imshow(encoded[j].reshape([1, 16]))\n",
    "    plot_tmp.axes.get_xaxis().set_visible(False)\n",
    "    plot_tmp.axes.get_yaxis().set_visible(False)\n",
    "    \n",
    "    fig.add_subplot(n_rows, 6, 3*j+3)\n",
    "    plot_tmp = imshow(decoded[j].reshape([28, 28]))\n",
    "    plot_tmp.axes.get_xaxis().set_visible(False)\n",
    "    plot_tmp.axes.get_yaxis().set_visible(False)\n",
    "    \n",
    "show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = read_csv(\"data/bank-additional-full.csv\", sep=\";\")\n",
    "y = (df.y == \"yes\") + 0\n",
    "del df[\"duration\"]\n",
    "del df[\"y\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.11265417111780131"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(y) / len(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>job</th>\n",
       "      <th>marital</th>\n",
       "      <th>education</th>\n",
       "      <th>default</th>\n",
       "      <th>housing</th>\n",
       "      <th>loan</th>\n",
       "      <th>contact</th>\n",
       "      <th>month</th>\n",
       "      <th>day_of_week</th>\n",
       "      <th>campaign</th>\n",
       "      <th>pdays</th>\n",
       "      <th>previous</th>\n",
       "      <th>poutcome</th>\n",
       "      <th>emp.var.rate</th>\n",
       "      <th>cons.price.idx</th>\n",
       "      <th>cons.conf.idx</th>\n",
       "      <th>euribor3m</th>\n",
       "      <th>nr.employed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>16533</th>\n",
       "      <td>31</td>\n",
       "      <td>services</td>\n",
       "      <td>divorced</td>\n",
       "      <td>high.school</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>telephone</td>\n",
       "      <td>jul</td>\n",
       "      <td>wed</td>\n",
       "      <td>2</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>1.4</td>\n",
       "      <td>93.918</td>\n",
       "      <td>-42.7</td>\n",
       "      <td>4.963</td>\n",
       "      <td>5228.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25956</th>\n",
       "      <td>39</td>\n",
       "      <td>blue-collar</td>\n",
       "      <td>married</td>\n",
       "      <td>basic.9y</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>cellular</td>\n",
       "      <td>nov</td>\n",
       "      <td>wed</td>\n",
       "      <td>1</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>-0.1</td>\n",
       "      <td>93.200</td>\n",
       "      <td>-42.0</td>\n",
       "      <td>4.120</td>\n",
       "      <td>5195.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9430</th>\n",
       "      <td>29</td>\n",
       "      <td>technician</td>\n",
       "      <td>married</td>\n",
       "      <td>university.degree</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>telephone</td>\n",
       "      <td>jun</td>\n",
       "      <td>fri</td>\n",
       "      <td>4</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>1.4</td>\n",
       "      <td>94.465</td>\n",
       "      <td>-41.8</td>\n",
       "      <td>4.967</td>\n",
       "      <td>5228.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25339</th>\n",
       "      <td>45</td>\n",
       "      <td>entrepreneur</td>\n",
       "      <td>married</td>\n",
       "      <td>basic.4y</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>cellular</td>\n",
       "      <td>nov</td>\n",
       "      <td>tue</td>\n",
       "      <td>5</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>-0.1</td>\n",
       "      <td>93.200</td>\n",
       "      <td>-42.0</td>\n",
       "      <td>4.153</td>\n",
       "      <td>5195.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39163</th>\n",
       "      <td>36</td>\n",
       "      <td>management</td>\n",
       "      <td>single</td>\n",
       "      <td>university.degree</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>telephone</td>\n",
       "      <td>mar</td>\n",
       "      <td>wed</td>\n",
       "      <td>1</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>-1.8</td>\n",
       "      <td>93.369</td>\n",
       "      <td>-34.8</td>\n",
       "      <td>0.655</td>\n",
       "      <td>5008.7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age           job   marital          education  default housing loan  \\\n",
       "16533   31      services  divorced        high.school       no      no   no   \n",
       "25956   39   blue-collar   married           basic.9y  unknown      no  yes   \n",
       "9430    29    technician   married  university.degree       no      no  yes   \n",
       "25339   45  entrepreneur   married           basic.4y       no     yes   no   \n",
       "39163   36    management    single  university.degree       no      no   no   \n",
       "\n",
       "         contact month day_of_week  campaign  pdays  previous     poutcome  \\\n",
       "16533  telephone   jul         wed         2    999         0  nonexistent   \n",
       "25956   cellular   nov         wed         1    999         0  nonexistent   \n",
       "9430   telephone   jun         fri         4    999         0  nonexistent   \n",
       "25339   cellular   nov         tue         5    999         0  nonexistent   \n",
       "39163  telephone   mar         wed         1    999         0  nonexistent   \n",
       "\n",
       "       emp.var.rate  cons.price.idx  cons.conf.idx  euribor3m  nr.employed  \n",
       "16533           1.4          93.918          -42.7      4.963       5228.1  \n",
       "25956          -0.1          93.200          -42.0      4.120       5195.8  \n",
       "9430            1.4          94.465          -41.8      4.967       5228.1  \n",
       "25339          -0.1          93.200          -42.0      4.153       5195.8  \n",
       "39163          -1.8          93.369          -34.8      0.655       5008.7  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = get_dummies(df)\n",
    "array_columns = X.columns\n",
    "min_max_scaler = MinMaxScaler()\n",
    "X = min_max_scaler.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "      <th>34</th>\n",
       "      <th>35</th>\n",
       "      <th>36</th>\n",
       "      <th>37</th>\n",
       "      <th>38</th>\n",
       "      <th>39</th>\n",
       "      <th>40</th>\n",
       "      <th>41</th>\n",
       "      <th>42</th>\n",
       "      <th>43</th>\n",
       "      <th>44</th>\n",
       "      <th>45</th>\n",
       "      <th>46</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>50</th>\n",
       "      <th>51</th>\n",
       "      <th>52</th>\n",
       "      <th>53</th>\n",
       "      <th>54</th>\n",
       "      <th>55</th>\n",
       "      <th>56</th>\n",
       "      <th>57</th>\n",
       "      <th>58</th>\n",
       "      <th>59</th>\n",
       "      <th>60</th>\n",
       "      <th>61</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "      <td>41188.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.284248</td>\n",
       "      <td>0.028502</td>\n",
       "      <td>0.963439</td>\n",
       "      <td>0.024709</td>\n",
       "      <td>0.725393</td>\n",
       "      <td>0.535723</td>\n",
       "      <td>0.430854</td>\n",
       "      <td>0.677237</td>\n",
       "      <td>0.769134</td>\n",
       "      <td>0.253035</td>\n",
       "      <td>0.224677</td>\n",
       "      <td>0.035350</td>\n",
       "      <td>0.025736</td>\n",
       "      <td>0.070992</td>\n",
       "      <td>0.041760</td>\n",
       "      <td>0.034500</td>\n",
       "      <td>0.096363</td>\n",
       "      <td>0.021244</td>\n",
       "      <td>0.163713</td>\n",
       "      <td>0.024619</td>\n",
       "      <td>0.008012</td>\n",
       "      <td>0.111974</td>\n",
       "      <td>0.605225</td>\n",
       "      <td>0.280859</td>\n",
       "      <td>0.001942</td>\n",
       "      <td>0.101389</td>\n",
       "      <td>0.055647</td>\n",
       "      <td>0.146766</td>\n",
       "      <td>0.231014</td>\n",
       "      <td>0.000437</td>\n",
       "      <td>0.127294</td>\n",
       "      <td>0.295426</td>\n",
       "      <td>0.042027</td>\n",
       "      <td>0.791201</td>\n",
       "      <td>0.208726</td>\n",
       "      <td>0.000073</td>\n",
       "      <td>0.452122</td>\n",
       "      <td>0.024036</td>\n",
       "      <td>0.523842</td>\n",
       "      <td>0.824269</td>\n",
       "      <td>0.024036</td>\n",
       "      <td>0.151695</td>\n",
       "      <td>0.634748</td>\n",
       "      <td>0.365252</td>\n",
       "      <td>0.063902</td>\n",
       "      <td>0.149995</td>\n",
       "      <td>0.004419</td>\n",
       "      <td>0.174177</td>\n",
       "      <td>0.129115</td>\n",
       "      <td>0.013256</td>\n",
       "      <td>0.334296</td>\n",
       "      <td>0.099568</td>\n",
       "      <td>0.017432</td>\n",
       "      <td>0.013839</td>\n",
       "      <td>0.190031</td>\n",
       "      <td>0.206711</td>\n",
       "      <td>0.209357</td>\n",
       "      <td>0.196416</td>\n",
       "      <td>0.197485</td>\n",
       "      <td>0.103234</td>\n",
       "      <td>0.863431</td>\n",
       "      <td>0.033335</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.128657</td>\n",
       "      <td>0.050364</td>\n",
       "      <td>0.187098</td>\n",
       "      <td>0.070700</td>\n",
       "      <td>0.327283</td>\n",
       "      <td>0.225581</td>\n",
       "      <td>0.193648</td>\n",
       "      <td>0.393210</td>\n",
       "      <td>0.273163</td>\n",
       "      <td>0.434756</td>\n",
       "      <td>0.417375</td>\n",
       "      <td>0.184665</td>\n",
       "      <td>0.158348</td>\n",
       "      <td>0.256814</td>\n",
       "      <td>0.200042</td>\n",
       "      <td>0.182513</td>\n",
       "      <td>0.295092</td>\n",
       "      <td>0.144199</td>\n",
       "      <td>0.370019</td>\n",
       "      <td>0.154962</td>\n",
       "      <td>0.089152</td>\n",
       "      <td>0.315339</td>\n",
       "      <td>0.488808</td>\n",
       "      <td>0.449424</td>\n",
       "      <td>0.044029</td>\n",
       "      <td>0.301846</td>\n",
       "      <td>0.229242</td>\n",
       "      <td>0.353877</td>\n",
       "      <td>0.421486</td>\n",
       "      <td>0.020901</td>\n",
       "      <td>0.333306</td>\n",
       "      <td>0.456239</td>\n",
       "      <td>0.200653</td>\n",
       "      <td>0.406455</td>\n",
       "      <td>0.406403</td>\n",
       "      <td>0.008534</td>\n",
       "      <td>0.497708</td>\n",
       "      <td>0.153163</td>\n",
       "      <td>0.499437</td>\n",
       "      <td>0.380596</td>\n",
       "      <td>0.153163</td>\n",
       "      <td>0.358729</td>\n",
       "      <td>0.481507</td>\n",
       "      <td>0.481507</td>\n",
       "      <td>0.244581</td>\n",
       "      <td>0.357071</td>\n",
       "      <td>0.066328</td>\n",
       "      <td>0.379266</td>\n",
       "      <td>0.335332</td>\n",
       "      <td>0.114372</td>\n",
       "      <td>0.471750</td>\n",
       "      <td>0.299427</td>\n",
       "      <td>0.130877</td>\n",
       "      <td>0.116824</td>\n",
       "      <td>0.392330</td>\n",
       "      <td>0.404951</td>\n",
       "      <td>0.406855</td>\n",
       "      <td>0.397292</td>\n",
       "      <td>0.398106</td>\n",
       "      <td>0.304268</td>\n",
       "      <td>0.343396</td>\n",
       "      <td>0.179512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.185185</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.340608</td>\n",
       "      <td>0.338912</td>\n",
       "      <td>0.160961</td>\n",
       "      <td>0.512287</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.259259</td>\n",
       "      <td>0.018182</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.937500</td>\n",
       "      <td>0.603274</td>\n",
       "      <td>0.376569</td>\n",
       "      <td>0.957379</td>\n",
       "      <td>0.859735</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.370370</td>\n",
       "      <td>0.036364</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.698753</td>\n",
       "      <td>0.602510</td>\n",
       "      <td>0.980957</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 0             1             2             3             4   \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.284248      0.028502      0.963439      0.024709      0.725393   \n",
       "std        0.128657      0.050364      0.187098      0.070700      0.327283   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.185185      0.000000      1.000000      0.000000      0.333333   \n",
       "50%        0.259259      0.018182      1.000000      0.000000      0.937500   \n",
       "75%        0.370370      0.036364      1.000000      0.000000      1.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 5             6             7             8             9   \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.535723      0.430854      0.677237      0.769134      0.253035   \n",
       "std        0.225581      0.193648      0.393210      0.273163      0.434756   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.340608      0.338912      0.160961      0.512287      0.000000   \n",
       "50%        0.603274      0.376569      0.957379      0.859735      0.000000   \n",
       "75%        0.698753      0.602510      0.980957      1.000000      1.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 10            11            12            13            14  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.224677      0.035350      0.025736      0.070992      0.041760   \n",
       "std        0.417375      0.184665      0.158348      0.256814      0.200042   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "75%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 15            16            17            18            19  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.034500      0.096363      0.021244      0.163713      0.024619   \n",
       "std        0.182513      0.295092      0.144199      0.370019      0.154962   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "75%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 20            21            22            23            24  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.008012      0.111974      0.605225      0.280859      0.001942   \n",
       "std        0.089152      0.315339      0.488808      0.449424      0.044029   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      1.000000      0.000000      0.000000   \n",
       "75%        0.000000      0.000000      1.000000      1.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 25            26            27            28            29  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.101389      0.055647      0.146766      0.231014      0.000437   \n",
       "std        0.301846      0.229242      0.353877      0.421486      0.020901   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "75%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 30            31            32            33            34  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.127294      0.295426      0.042027      0.791201      0.208726   \n",
       "std        0.333306      0.456239      0.200653      0.406455      0.406403   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      1.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      1.000000      0.000000   \n",
       "75%        0.000000      1.000000      0.000000      1.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 35            36            37            38            39  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.000073      0.452122      0.024036      0.523842      0.824269   \n",
       "std        0.008534      0.497708      0.153163      0.499437      0.380596   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      1.000000   \n",
       "50%        0.000000      0.000000      0.000000      1.000000      1.000000   \n",
       "75%        0.000000      1.000000      0.000000      1.000000      1.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 40            41            42            43            44  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.024036      0.151695      0.634748      0.365252      0.063902   \n",
       "std        0.153163      0.358729      0.481507      0.481507      0.244581   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      1.000000      0.000000      0.000000   \n",
       "75%        0.000000      0.000000      1.000000      1.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 45            46            47            48            49  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.149995      0.004419      0.174177      0.129115      0.013256   \n",
       "std        0.357071      0.066328      0.379266      0.335332      0.114372   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "75%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 50            51            52            53            54  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.334296      0.099568      0.017432      0.013839      0.190031   \n",
       "std        0.471750      0.299427      0.130877      0.116824      0.392330   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "75%        1.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 55            56            57            58            59  \\\n",
       "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
       "mean       0.206711      0.209357      0.196416      0.197485      0.103234   \n",
       "std        0.404951      0.406855      0.397292      0.398106      0.304268   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "75%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "                 60            61  \n",
       "count  41188.000000  41188.000000  \n",
       "mean       0.863431      0.033335  \n",
       "std        0.343396      0.179512  \n",
       "min        0.000000      0.000000  \n",
       "25%        1.000000      0.000000  \n",
       "50%        1.000000      0.000000  \n",
       "75%        1.000000      0.000000  \n",
       "max        1.000000      1.000000  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DataFrame(X).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = RandomForestClassifier(\n",
    "    n_estimators=500, \n",
    "    max_depth=2, \n",
    "    n_jobs=8, \n",
    "    random_state=42\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "dict_performance = cross_validate(\n",
    "    estimator=rf, \n",
    "    X=X, y=y, \n",
    "    cv=10, \n",
    "    n_jobs=4,\n",
    "    return_train_score=True,\n",
    "    scoring=[\n",
    "        \"balanced_accuracy\", \n",
    "        \"f1_weighted\", \n",
    "        \"roc_auc\", \n",
    "        \"average_precision\"\n",
    "    ]\n",
    ")\n",
    "\n",
    "df_performance = DataFrame(\n",
    "    {\"ORIGINAL\": [mean(dict_performance[k]) \\\n",
    "                  for k in dict_performance.keys()]}, \n",
    "    index=dict_performance.keys()\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>ORIGINAL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>fit_time</th>\n",
       "      <td>8.061133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>score_time</th>\n",
       "      <td>0.830988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_balanced_accuracy</th>\n",
       "      <td>0.505833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_balanced_accuracy</th>\n",
       "      <td>0.517641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_f1_weighted</th>\n",
       "      <td>0.809805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_f1_weighted</th>\n",
       "      <td>0.842227</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_roc_auc</th>\n",
       "      <td>0.563038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_roc_auc</th>\n",
       "      <td>0.801313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_average_precision</th>\n",
       "      <td>0.414990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_average_precision</th>\n",
       "      <td>0.463692</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         ORIGINAL\n",
       "fit_time                 8.061133\n",
       "score_time               0.830988\n",
       "test_balanced_accuracy   0.505833\n",
       "train_balanced_accuracy  0.517641\n",
       "test_f1_weighted         0.809805\n",
       "train_f1_weighted        0.842227\n",
       "test_roc_auc             0.563038\n",
       "train_roc_auc            0.801313\n",
       "test_average_precision   0.414990\n",
       "train_average_precision  0.463692"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBEAAAHwCAYAAADw2/frAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/d3fzzAAAACXBIWXMAAAsTAAALEwEAmpwYAAA8ZklEQVR4nO3de5hdVX3/8fcHolwChCKKgEIUjFwUUVMURUVJba0ioKJFq1KslGK91FJLTeWHtVittlZrUZEqKFgRKorQclVuUQmESxJuooJFpVqkchUk5Pv746yBwziZ7ISZnDOT9+t5zjP7rL332t+9l3gyn1l7n1QVkiRJkiRJK7POoAuQJEmSJElTgyGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJIkiRJkqRODBEkSRJJ3pvk2InetkNflWT7FazbIsmFSe5M8o8TcbxhkOTqJHsOug5JklZHqmrQNUiSpAmU5EDgL4DtgDuAU4G/rqpfDrCsMSUp4ClV9f0x1r0PeCbw6nqE/2BJchzw46r6m0fSz3ST5Cbgj6vq3EHXIkmaGpyJIEnSNJLkL4APA38JzAKeC2wLnJPk0SvYZ8aaq3CVbAtc80gDhIkwxNdotUy385EkrTmGCJIkTRNJNgHeD7y9qs6sqvur6ibgtfR+If/Dtt2RSU5JckKSO4ADW9sJfX29KcmPkvwiyfuS3JRkXt/+J7Tl2e2WhDcn+e8ktyaZ39fPbkm+k+SXSW5J8skVhRmjzuU44M3Ae5LclWReknWSHJ7kB62uryTZrG+fk5P8T5Lb220QO7f2g4E39PX1jdb+sFspkhyX5O/a8p5Jfpzkr5L8D/D58Y6fZP12PX/RzvXSJFus4NxGX8uT2753JlmSZE6Sv07y8yQ3J3lp377nJ/n7JAvbeX591DV4Zbtd4pdt2x1HHfevkiwG7k7y78A2wDfadXnPeNex7xr9a5IzWr2XJNmub/3OSc5JcluSnyV5b2sfd+wkSVOHIYIkSdPH84D1ga/2N1bVXcB/Ab/T17wPcAqwKXBi//ZJdgKOpveL95b0ZjRsvZJj7wE8FdgLOKLvl9cHgD8HNgd2b+sPXdmJVNWBra5/qKqN2nT7dwD7Ai8CtgL+D/jXvt3+C3gK8Djg8pHzqqpjRvW198qO3zwe2IxeAHPwSo7/ZnrX6YnAY4BDgF91PM7ewBeB3wKuAM6i92+0rYG/BT4zavs3AQe1GpYBnwBIMgf4d+BdwGOB/6QXEPSHNgcALwc2raoDgP8G9m7X5R/aNmNex1F9vL/V+33gqHb8jYFzgTNbbdsD57V9VjZ2kqQpwhBBkqTpY3Pg1qpaNsa6W9r6Ed+pqq9V1fKqGv3L7muAb1TVxVX1a+AIYGW3FLy/qn5VVVcBVwHPAKiqRVX13apa1mZFfIbeL5Kr40+A+VX146q6DzgSeM3I1Pyq+lxV3dm37hlJZq3msQCWA/+vqu5r12i8499PLzzYvqoeaOd9R8fjXFRVZ7VxO5leAPChqrof+DIwO8mmfdt/saqWVtXdwPuA1yZZF3gdcEZVndP2/SiwAb1wacQnqurmMcb8QR2u41eramGr90Rg19b+CuB/quofq+re1sclbd24YydJmjr8P25JkqaPW4HNk8wYI0jYsq0fcfM4/WzVv76q7knyi5Uc+3/6lu8BNoIH/zr+T8BcYEN6//ZYtJK+VmRb4NQky/vaHgC2aLccHAXsT++X8JFtNgduX83j/W9V3dvl+PRmEjwR+HL7hf8Eer8039/hOD/rW/4VvSDogb730Luev2zL/WP3I+BR9M5zq/YegKpanuRmHj6LZLxxp4URK7uOY441vfP/wQq6Hu/a/WS8miRJw8WZCJIkTR/fAe4DXtXfmGQm8DIemloO488suAV4Qt/+G9D7K/vq+BRwHb1vYNgEeC+Q1ezrZuBlVbVp32v9qvoJ8Hp6t2jMo3dbweyR8tvPsc73HnrBxojHj1o/ep8VHr89f+L9VbUTvb/8v4LebQeT4Yl9y9vQmwVxK/BTer+sA5Akbdv+X9JHn9Po9yu7juO5md43gqxo3YrGTpI0hRgiSJI0TVTV7fTuVf+XJL+X5FFJZtObIv9jen8t7+IUYO8kz2v307+f1f/Ff2N6XzN5V5IdgD9dzX4APg0clWRbgCSPTbJP33HuA35BLxj44Kh9fwY8eVTblcDrk6yb5PdY+W0WKzx+khcneXr7S/4d9H6xf2DFXT0if5hkpyQb0ntmwilt5sJXgJcn2SvJo+h9zed9wLfH6Wv0dVnZdRzP6cDjk7wryXpJNk7ynLZuvLGTJE0hhgiSJE0j7eF476V3P/wdwCX0/gq8V7sXvUsfVwNvp3c//i3AncDP6f1yuaoOo/fX7TuBzwInrUYfIz4OnAacneRO4LvAyC+pX6A3lf8nwDVtXb9/A3Zq31rwtdb2TnoPNfwlvYdIfo3xjXf8x9MLX+4ArgUuoHdLw2T4InAcvdsK1qf30EKq6np638DxL/RmJuxN76GJvx6nr78H/qZdl8NY+XVcoaq6k97DO/dutd0AvLitHu/aSZKmkAzBVy9LkqQhlmTkfvynVNWNAy5nrZbkfOCEqjp20LVIktZOzkSQJEm/IcneSTZsz1P4KLAEuGmwVUmSpEEzRJAkSWPZh96D+n4KPAX4g3L6oiRJaz1vZ5AkSZIkSZ04E0GSJEmSJHViiCBJkiRJkjqZMegCNDVtvvnmNXv27EGXIUmSJEmaBIsWLbq1qh47ut0QQatl9uzZXHbZZYMuQ5IkSZI0CZL8aKx2b2eQJEmSJEmdGCJIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSerEEEGSJEmSJHViiCBJkiRJkjoxRJAkSZIkSZ0YIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTmYMugBNTdfecge7HXXuoMuQJEmSpClj4fx5gy7hEXMmgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSYRpL8Z5JNB12HJEmSJGl6mjHoAjS2JOtW1QOrsk9V/f5k1SNJkiRJkjMRBiDJ7CTXJTk+yeIkpyTZMMlNSY5IcjGwf5KXJvlOksuTnJxkoyQvS/KVvr72TPKNtnxTks3b8ruTLG2vd/Udd2nfvoclObItvyPJNa2eL6/ByyFJkiRJmiKciTA4TwXeUlULknwOOLS131tVe7Qw4KvAvKq6O8lfAe8GPgh8JsnMqrobeB1wUn/HSZ4N/BHwHCDAJUkuAP5vnHoOB55UVfd5S4QkSZIkaSzORBicm6tqQVs+AdijLY8EAs8FdgIWJLkSeDOwbVUtA84E9k4yA3g58PVRfe8BnFpVd1fVXfTCiBespJ7FwIlJ/hBYNtYGSQ5OclmSy5bdfXvX85QkSZIkTRPORBicWsH7u9vPAOdU1QFj7HsS8DbgNuDSqrpz1Pqs4JjLeHhwtH7f8suBFwKvBN6XZOcWWDxUYNUxwDEAM7eeM7p+SZIkSdI050yEwdkmye5t+QDg4lHrvws8P8n2AO2ZCXPauvOBZwFvZdStDM2FwL5tn5nAfsBFwM+AxyV5TJL1gFe0vtcBnlhV3wLeA2wKbDQhZylJkiRJmjYMEQbnWuDNSRYDmwGf6l9ZVf8LHAj8e9vmu8AObd0DwOnAy9pPRu17OXAcsBC4BDi2qq6oqvuBv21tpwPXtV3WBU5IsgS4AvhYVf1yAs9VkiRJkjQNpMpZ6WtaktnA6VX1tEHXsrpmbj2ndj706EGXIUmSJElTxsL58wZdQmdJFlXV3NHtzkSQJEmSJEmd+GDFAaiqm4ApOwtBkiRJkrR2ciaCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSepkxqAL0NS045absHD+vEGXIUmSJElag5yJIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqZMagC9DUdO0td7DbUecOugxJkiRJmjAL588bdAlDz5kIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSepkrQsRkuyZ5HmDrkOSJEmSpKlmrQsRgD0BQwRJkiRJklbRlAkRksxOcl2S45MsTnJKkg2T7JXkiiRLknwuyXpt+5uSbN6W5yY5P8ls4BDgz5NcmeQFSbZIcmqSq9rreW2fdydZ2l7vGlXDsa39xCTzkixIckOS3dp2M1stl7ba9hnnvHZOsrDVszjJU9pxlvZtc1iSI9vy9knObbVenmS71v6edg2uSvKh1rZdkjOTLEpyUZIdWvv+rf6rkly4ojomdAAlSZIkSVPejEEXsIqeCrylqhYk+RzwbuBPgL2q6ntJvgD8KfDPY+1cVTcl+TRwV1V9FCDJScAFVbVfknWBjZI8G/gj4DlAgEuSXAD8H7A9sD9wMHAp8HpgD+CVwHuBfYH5wDer6qAkmwILk5xbVXePUdYhwMer6sQkjwbWBbYY5xqcCHyoqk5Nsj6wTpKXteM+p6ruSbJZ2/YY4JCquiHJc4CjgZcARwC/W1U/afWtqA5JkiRJkh40ZWYiNDdX1YK2fAKwF3BjVX2vtR0PvHAV+3wJ8CmAqnqgqm6nFwqcWlV3V9VdwFeBF7Ttb6yqJVW1HLgaOK+qClgCzG7bvBQ4PMmVwPnA+sA2Kzj+d4D3JvkrYNuq+tWKCk2yMbB1VZ3a6r23qu4B5gGfb8tU1W1JNqJ328bJrY7PAFu2rhYAxyV5Kw+FBSutI8nBSS5Lctmyu29fUZmSJEmSpGlqqoUItQrbLuOh81t/FY+Tcdbd17e8vO/9ch6a2RHg1VW1a3ttU1XXjtVZVX2J3iyGXwFnJXnJqNr7619RXeE3r806wC/7ati1qnZsxzwE+BvgicCVSR6zgjpG13pMVc2tqrkzZs5aQSmSJEmSpOlqqoUI2yTZvS0fAJwLzE6yfWt7I3BBW74JeHZbfnVfH3cCG/e9P4/eLRAkWTfJJsCFwL7tmQszgf2Ai1ahzrOAtydJ6/eZK9owyZOBH1bVJ4DTgF2AnwGPS/KY9oyHVwBU1R3Aj5Ps2/ZdL8mGwNnAQW2ZJJu1bW9Msn9rS5JntOXtquqSqjoCuBV44grqkCRJkiTpQVMtRLgWeHOSxcBmwMfoPbvg5CRL6M0G+HTb9v3Ax5NcBDzQ18c3gP1GHqwIvBN4cdt/EbBzVV0OHAcsBC4Bjq2qK1ahzg8AjwIWtwckfmCcbV8HLG23HOwAfKGq7gf+th37dOC6vu3fCLyjXYNvA4+vqjPp/eJ/WevnsLbtG4C3JLmK3q0XIw94/Eh7CONSeoHJVWPVsQrnK0mSJElaC6R3O//wS++bFU6vqqcNuhbBzK3n1M6HHj3oMiRJkiRpwiycP2/QJQyNJIuqau7o9qk2E0GSJEmSJA3IlPmKx6q6CZiysxCS/C7w4VHNN1bVfoOoR5IkSZKkVTVlQoSprqrOovfARUmSJEmSpiRvZ5AkSZIkSZ0YIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOpkx6AI0Ne245SYsnD9v0GVIkiRJktYgZyJIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSerEEEGSJEmSJHViiCBJkiRJkjqZMegCNDVde8sd7HbUuYMuQ5IkSZImzML58wZdwtBzJoIkSZIkSerEEEGSJEmSJHViiCBJkiRJkjoxRJAkSZIkSZ0YIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiTCNJ9kxy+qDrkCRJkiRNT4YIkiRJkiSpE0OEKSLJ7CTXJTk+yeIkpyTZMMnvtfaLgVf1bb9bkm8nuaL9fGprvyjJrn3bLUiyS5IXJbmyva5IsvGaP0tJkiRJ0jAzRJhangocU1W7AHcA7wY+C+wNvAB4fN+21wEvrKpnAkcAH2ztxwIHAiSZA6xXVYuBw4C3VdWura9fTfbJSJIkSZKmFkOEqeXmqlrQlk8A5gI3VtUNVVWtbcQs4OQkS4GPATu39pOBVyR5FHAQcFxrXwD8U5J3AJtW1bLRB09ycJLLkly27O7bJ/rcJEmSJElDzhBhaqlR72eN0TbiA8C3qupp9GYqrA9QVfcA5wD7AK8FvtTaPwT8MbAB8N0kO/zGwauOqaq5VTV3xsxZE3A6kiRJkqSpxBBhatkmye5t+QDgXOBJSbbraxsxC/hJWz5wVD/HAp8ALq2q2wCSbFdVS6rqw8BlwG+ECJIkSZKktZshwtRyLfDmJIuBzejdpnAwcEZ7sOKP+rb9B+DvkywA1u3vpKoW0Xumwuf7mt+VZGmSq+g9D+G/Ju80JEmSJElT0YxBF6BVsryqDhnVdiZjzBqoqu8Ac/qa3jeykGQregHS2X3bv31iS5UkSZIkTTfORFjLJHkTcAkwv6qWD7oeSZIkSdLU4UyEKaKqbgKeNgH9fAH4wiMuSJIkSZK01nEmgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqZMagC9DUtOOWm7Bw/rxBlyFJkiRJWoOciSBJkiRJkjoxRJAkSZIkSZ0YIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6mTGoAvQ1HTtLXew21HnDroMSZIkSUNi4fx5gy5Ba4AzESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSHCkEnyyiSHt+XjkrzmEfS1bZJFSa5McnWSQyauUkmSJEnS2mbGoAvQQ5LMqKrTgNMmoi/gFuB5VXVfko2ApUlOq6qfPtL+JUmSJElrH2ciTIIkf5hkYZsB8Jkk6ya5q2/9a5Ic15aPS/JPSb4FfDjJgUk+2dfdvCQXJflekle0fdZP8vkkS5JckeTFrf3AJCcn+QZwdlX9uqrua/2sR994J7kpyQeTfCfJZUmeleSsJD9wxoIkSZIkaSzORJhgSXYEXgc8v6ruT3I08IaV7DYHmFdVDyQ5cNS62cCLgO2AbyXZHngbQFU9PckOwNlJ5rTtdwd2qarbWj1PBM4Atgf+ctQshJuravckHwOOA54PrA9cDXx6lU9ekiRJkjStGSJMvL2AZwOXJgHYAPj5SvY5uaoeWMG6r1TVcuCGJD8EdgD2AP4FoKquS/IjekEEwDkjAUJbfzOwS5KtgK8lOaWqftZWj9w2sQTYqKruBO5Mcm+STavql/2FJDkYOBjg0bMet5JTkiRJkiRNN97OMPECHF9Vu7bXU6vqSKD6tll/1D53j9NfjfE+42w/Zl9tBsLVwAv6mkdudVjetzzy/jcCpqo6pqrmVtXcGTNnjVOCJEmSJGk6MkSYeOcBr0nyOIAkmyXZFvhZkh2TrAPstwr97Z9knSTbAU8GrgcupN0i0W5j2Ka1P0ySJyTZoC3/Fr3bFX5jO0mSJEmSuvB2hglWVdck+Rt6zylYB7if3jMMDgdOB24GlgIbdezyeuACYAvgkKq6tz1n4dNJlgDLgAPbNzCM3ndH4B+TjMxe+GhVLXlkZyhJkiRJWlulavRseWnlZm49p3Y+9OhBlyFJkiRpSCycP2/QJWgCJVlUVXNHt3s7gyRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6mTHoAjQ17bjlJiycP2/QZUiSJEmS1iBnIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOpkx6AI0NV17yx3sdtS5gy5DkiRJ0pBYOH/eoEvQGuBMBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJMU0m+lmRRkquTHNza3pLke0nOT/LZJJ9s7Y9N8h9JLm2v5w+2ekmSJEnSMJox6AI0aQ6qqtuSbABcmuQM4H3As4A7gW8CV7VtPw58rKouTrINcBaw4yCKliRJkiQNL0OE6esdSfZry08E3ghcUFW3ASQ5GZjT1s8Ddkoysu8mSTauqjv7O2wzGg4GePSsx01y+ZIkSZKkYWOIMA0l2ZNeMLB7Vd2T5HzgelY8u2Cdtu2vxuu3qo4BjgGYufWcmqh6JUmSJElTg89EmJ5mAf/XAoQdgOcCGwIvSvJbSWYAr+7b/mzgz0beJNl1TRYrSZIkSZoaDBGmpzOBGUkWAx8Avgv8BPggcAlwLnANcHvb/h3A3CSLk1wDHLLmS5YkSZIkDbvOtzO0B/RtU1XXT2I9mgBVdR/wstHtSS6rqmPaTIRT6c1AoKpuBV63ZquUJEmSJE01nWYiJNkbuJLeX7hJsmuS0yaxLk2OI5NcCSwFbgS+NtBqJEmSJElTSteZCEcCuwHnA1TVlUlmT05JmixVddiga5AkSZIkTV1dn4mwrKpuX/lmkiRJkiRpuuo6E2FpktcD6yZ5Cr0H8X178sqSJEmSJEnDputMhLcDOwP3AV+i91T/d01STZIkSZIkaQitdCZCknWB06pqHjB/8kuSJEmSJEnDaKUzEarqAeCeJLPWQD2SJEmSJGlIdX0mwr3AkiTnAHePNFbVOyalKkmSJEmSNHS6hghntJckSZIkSVpLdQoRqur4yS5EkiRJkiQNt04hQpIbgRrdXlVPnvCKJEmSJEnSUOp6O8PcvuX1gf2BzSa+HE0VO265CQvnzxt0GZIkSZKkNWil384AUFW/6Hv9pKr+GXjJ5JYmSZIkSZKGSdfbGZ7V93YdejMTNp6UiiRJkiRJ0lDqejvDP/YtLwNuBF478eVIkiRJkqRh1TVEeEtV/bC/IcmTJqEeSZIkSZI0pDo9EwE4pWObJEmSJEmapsadiZBkB2BnYFaSV/Wt2oTetzRIkiRJkqS1xMpuZ3gq8ApgU2DvvvY7gbdOUk2SJEmSJGkIpapWvlGye1V9Zw3Uoyli5tZzaudDjx50GZIkSZI6Wjh/3qBL0BSSZFFVzR3d3vXBilckeRu9WxsevI2hqg6aoPokSZIkSdKQ6/pgxS8Cjwd+F7gAeAK9WxokSZIkSdJaomuIsH1VvQ+4u6qOB14OPH3yypIkSZIkScOma4hwf/v5yyRPA2YBsyelIkmSJEmSNJS6PhPhmCS/BbwPOA3YCDhi0qqSJEmSJElDp1OIUFXHtsULgCdPXjmSJEmSJGlYdbqdIckWSf4tyX+19zslecvkliZJkiRJkoZJ12ciHAecBWzV3n8PeNck1CNJkiRJkoZU1xBh86r6CrAcoKqWAQ9MWlWSJEmSJGnodA0R7k7yGKAAkjwXuH3SqpIkSZIkSUOn67czvJvetzJsl2QB8FjgNZNWlSRJkiRJGjrjhghJtqmq/66qy5O8CHgqEOD6qrp/jVQoSZIkSZKGwspuZ/ha3/JJVXV1VS01QBicJLOTLJ3kY5yfZO5kHkOSJEmSNPWsLERI3/KTJ7OQ6SBJ19tDJEmSJEmaclYWItQKltc6bQbAtUk+m+TqJGcn2aD91f6DSS4A3jlqn79McmmSxUne39fPdUmOTbI0yYlJ5iVZkOSGJLu17Y5M8sUk32ztbx2jpvWTfD7JkiRXJHlxa78oya592y1IskuSmUk+12q6Isk+bf0GSb7c6jwJ2GDSLqQkSZIkacpa2V/On5HkDnozEjZoy7T3VVWbTGp1w+cpwAFV9dYkXwFe3do3raoX9W+Y5KVt+93oXa/TkrwQ+G9ge2B/4GDgUuD1wB7AK4H3Avu2bnYBngvMBK5Icsaoet4GUFVPT7IDcHaSOcCxwIHAu9r79apqcZIPAt+sqoOSbAosTHIu8CfAPVW1S5JdgMsf4XWSJEmSJE1D485EqKp1q2qTqtq4qma05ZH3a1uAAHBjVV3ZlhcBs9vySWNs+9L2uoLeL+U70AsVRvpZUlXLgauB86qqgCV9fQJ8vap+VVW3At+iF0j02wP4IkBVXQf8CJgDnAy8IsmjgIOA4/pqOjzJlcD5wPrANsALgRNaP4uBxWOdfJKDk1yW5LJld/sNn5IkSZK0tvEe/lVzX9/yAzw07f/uMbYN8PdV9ZmHNSazR/WzvO/9ch4+JqNvIRn9Poyhqu5Jcg6wD/BaYG7f9q+uqutH1TRW32P1ewxwDMDMrees1be3SJIkSdLaaGXPRNDqOws4KMlGAEm2TvK4Vexjn/bcg8cAe9K79aHfhcAbWv9z6M0qGAkIjgU+AVxaVbf11fT2tNQgyTPH6Odp9G6jkCRJkiTpYQwRJlCSuUmOBaiqs4EvAd9JsgQ4Bdh4FbtcCJwBfBf4QFX9dNT6o4F1W/8nAQdW1X3t+IuAO4DP923/AeBRwOL2NZEfaO2fAjZKshh4TzuuJEmSJEkPk96t+Bo2SY4E7qqqj67m/lvRe+7BDu3ZCxNq5tZzaudDj57obiVJkiRNkoXz5w26BE0hSRZV1dzR7c5EmIaSvAm4BJg/GQGCJEmSJGnt5IMVh1RVHfkI9v0C8IWJq0aSJEmSJGciSJIkSZKkjgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJIkiRJkqROZgy6AE1NO265CQvnzxt0GZIkSZKkNciZCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTmYMugBNTdfecge7HXXuoMuQJEmSpo2F8+cNugRppZyJIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJIkiRJkqRODBGmiSQvSHJ1kiuTbNDX/rdJ5o2x/Z5JTl+zVUqSJEmSprIZgy5AE+YNwEer6vP9jVV1xIDqkSRJkiRNM85EGEeSNyVZnOSqJF9Msm2S81rbeUm2adsdl+QTSb6d5IdJXtPat0xyYZsdsDTJC8Y4xvZJzm3HuDzJdun5SNtnSZLXtW33THJ+klOSXJfkxLbtHwOvBY5IcuKo/o/rq+f32n4XA6/q2+YTSY5oy7/bavZ/G5IkSZKkh3Emwgok2RmYDzy/qm5NshlwPPCFqjo+yUHAJ4B92y5bAnsAOwCnAacArwfOqqqjkqwLbDjGoU4EPlRVpyZZn16w8ypgV+AZwObApUkubNs/E9gZ+CmwoNV3bJI9gNOr6pQVnM/6wGeBlwDfB07qW314O8ZF7Zx+v6qWd79akiRJkqS1gX9tXrGXAKdU1a0AVXUbsDvwpbb+i/RCgxFfq6rlVXUNsEVruxT4oyRHAk+vqjv7D5BkY2Drqjq1HePeqrqn9fvvVfVAVf0MuAD47bbbwqr6cfsl/0pgdsfz2QG4sapuqKoCThhZ0Y75VuAc4JNV9YOxOkhycJLLkly27O7bOx5WkiRJkjRdGCKsWIBayTb96+8btS9VdSHwQuAnwBeTvGmMY6zo2CvSf5wHWLXZJOOdz9OBXwBbrXDnqmOqam5VzZ0xc9YqHFaSJEmSNB0YIqzYecBrkzwGoN3O8G3gD9r6NwAXj9dBkm2Bn1fVZ4F/A57Vv76q7gB+nGTftv16STYELgRel2TdJI+lF0QsfITncx3wpCTbtfcHjKrzL+jdKvGyJM95hMeSJEmSJE1DhggrUFVXA0cBFyS5Cvgn4B30bk9YDLwReOdKutkTuDLJFcCrgY8DJDk2ydy2zRuBd7Q+vw08HjgVWAxcBXwTeE9V/U/X2kf1P3I+9wIHA2e0Byv+qG0begHHYVX1U+AtwLHtGQqSJEmSJD0ovdvjpVUzc+s5tfOhRw+6DEmSJGnaWDh/3qBLkB6UZFFVzR3d7kwESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSerEEEGSJEmSJHUyY9AFaGracctNWDh/3qDLkCRJkiStQc5EkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1MmPQBWhquvaWO9jtqHMHXYYkSdKUtHD+vEGXIEmrxZkIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCJIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSerEEEGSJEmSJHViiDCkkmya5NC+93smOX01+1rtfSVJkiRJGmGIMLw2BQ5d2UaSJEmSJK0phggTIMnsJNclOTbJ0iQnJpmXZEGSG5LslmSzJF9LsjjJd5Ps0vY9Msnnkpyf5IdJ3tG6/RCwXZIrk3yktW2U5JR2rBOTZJyafq9tdzHwqr72me14lya5Isk+rX3dJB9NsqTV+PbJuVqSJEmSpKlqxqALmEa2B/YHDgYuBV4P7AG8EngvcDNwRVXtm+QlwBeAXdu+OwAvBjYGrk/yKeBw4GlVtSv0bkkAngnsDPwUWAA8H7h4dCFJ1gc+C7wE+D5wUt/q+cA3q+qgJJsCC5OcC7wJeBLwzKpalmSzMfo9uJ0fj571uFW8PJIkSZKkqc6ZCBPnxqpaUlXLgauB86qqgCXAbHqBwhcBquqbwGOSzGr7nlFV91XVrcDPgS1WcIyFVfXjdowrW79j2aHVc0Or4YS+dS8FDk9yJXA+sD6wDTAP+HRVLWs13ja606o6pqrmVtXcGTNnjV4tSZIkSZrmnIkwce7rW17e9345veu8bIx9aox9H2DF49J1u/6+Rwvw6qq6/mGNvVsjVrSPJEmSJEnORFiDLgTeAA/emnBrVd0xzvZ30ru9YXVcBzwpyXbt/QF9684C3j7yPIUkz2ztZwOHJJnR2n/jdgZJkiRJ0trNEGHNORKYm2QxvYcmvnm8javqF8CC9qDGj4y37Rj73kvv2QVntAcr/qhv9QeARwGLkyxt7wGOBf67tV9F75kOkiRJkiQ9KL1b5qVVM3PrObXzoUcPugxJkqQpaeH8eYMuQZLGlWRRVc0d3e5MBEmSJEmS1IkPVpzikpxK76sZ+/1VVZ01iHokSZIkSdOXIcIUV1X7DboGSZIkSdLawdsZJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdzBh0AZqadtxyExbOnzfoMiRJkiRJa5AzESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJncwYdAGamq695Q52O+rcQZchSZI0FBbOnzfoEiRpjXAmgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdGCKsZZJsmuTQQdchSZIkSZp6DBGmsCTrrqB9xji7bQoYIkiSJEmSVpkhwkok+cMkC5NcmeQzSdZNcleSDydZlOTcJLslOT/JD5O8su13YJKvJzkzyfVJ/t8Yfb8syVf63u+Z5Btt+VNJLktydZL3921zU5IjklwM7N/XfmCSk9v+ZyfZKMl5SS5PsiTJPm3TDwHbtfP5SNv3L5NcmmRx/7EkSZIkSeo33l+s13pJdgReBzy/qu5PcjTwBmAmcH5V/VWSU4G/A34H2Ak4HjitdbEb8DTgHuDSJGdU1WV9hzgH+EySmVV1dzvWSW3d/Kq6rc02OC/JLlW1uK27t6r2GKPk3YFd2n4zgP2q6o4kmwPfTXIacDjwtKratZ3jS4GntFoDnJbkhVV14SO5dpIkSZKk6ccQYXx7Ac+mFwAAbAD8HPg1cGbbZglwXwsZlgCz+/Y/p6p+AZDkq8AewIMhQlUtS3ImsHeSU4CXA+9pq1+b5GB6Y7QlvYBiJEQYCRpGO6eqbmvLAT6Y5IXAcmBrYIsx9nlpe13R3m9EL1T4jRCh1XMwwKNnPW4FJUiSJEmSpitDhPEFOL6q/vphjclhVVXt7XLgPoCqWj7qeQTFw41+D71A4G3AbcClVXVnkicBhwG/XVX/l+Q4YP2+fe5eQb397W8AHgs8uwUcN43qo/8c/76qPrOCPh8qvuoY4BiAmVvPGetcJEmSJEnTmM9EGN95wGuSPA4gyWZJtl2F/X+n7bMBsC+wYIxtzgeeBbyVh2YYbEIvELg9yRbAy1aj9lnAz1uA8GJgpO47gY37tjsLOCjJRgBJth45X0mSJEmS+hkijKOqrgH+ht6DChfTe4bBlqvQxcXAF4Ergf8YeR5Ckv9MslU7xgPA6fSCgtNb21X0bi+4GvgcY4cPJHllkr9dwbFPBOYmuYzerITrWt+/ABYkWZrkI1V1NvAl4DvtdoxTeHjIIEmSJEkSAHloVr4mUpIDgblV9WeDrmUyzNx6Tu186NGDLkOSJGkoLJw/b9AlSNKESrKoquaObncmgiRJkiRJ6sQHK06SqjoOOG7AZUiSJEmSNGGciSBJkiRJkjoxRJAkSZIkSZ0YIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOpkx6AI0Ne245SYsnD9v0GVIkiRJktYgZyJIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSerEEEGSJEmSJHViiCBJkiRJkjqZMegCNDVde8sd7HbUuYMuQ5IkqZOF8+cNugRJmhaciSBJkiRJkjoxRJAkSZIkSZ0YIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ0MfIiTZM8nzBl3HMEjyt0nmjbN+3yQ7PYL+Zyd5/eruL0mSJEma3oY+RAD2BAwRgKo6oqrOHWeTfYHVDhGA2YAhgiRJkiRpTJMWIrS/al+X5Pgki5OckmTDJHsluSLJkiSfS7Je2/6mJJu35blJzk8yGzgE+PMkVyZ5QZItkpya5Kr2el7b591JlrbXu0bVcGxrPzHJvCQLktyQZLe23cxWy6Wttn3GOa8Dk3w1yZmtj3/oW3dAO6+lST7c135XkqNavd9NskVrf2yS/2jHvTTJ81v715O8qS3/SZIT2/JxSV7Tlj+U5Jp2bT/arsMrgY+0a7Vde52ZZFGSi5Ls0NfPJ5J8O8kPR/oEPgS8oO3/54/sfwGSJEmSpOlmxiT3/1TgLVW1IMnngHcDfwLsVVXfS/IF4E+Bfx5r56q6Kcmngbuq6qMASU4CLqiq/ZKsC2yU5NnAHwHPAQJckuQC4P+A7YH9gYOBS+n9pX0Per9wv5feX+/nA9+sqoOSbAosTHJuVd29gvPaFXgmcB9wfZJ/AR4APgw8ux337CT7VtXXgJnAd6tqfgsd3gr8HfBx4GNVdXGSbYCzgB1brQuS3Aj8BfDc/oMn2QzYD9ihqirJplX1yySnAadX1Sltu/OAQ6rqhiTPAY4GXtK62bJdhx2A04BTgMOBw6rqFWOddJKDW208etbjVnBpJEmSJEnT1WTfznBzVS1oyycAewE3VtX3WtvxwAtXsc+XAJ8CqKoHqup2er8Mn1pVd1fVXcBXgRe07W+sqiVVtRy4GjivqgpYQm/6PsBLgcOTXAmcD6wPbDNODedV1e1VdS9wDbAt8NvA+VX1v1W1DDix79x+DZzelhf1HXce8Ml23NOATZJsXFU/A44AvgX8RVXdNur4dwD3AscmeRVwz+gCk2xE7zaQk1v/n6EXHIz4WlUtr6prgC3GOdcHVdUxVTW3qubOmDmryy6SJEmSpGlksmci1Cpsu4yHQo31V/E4GWfdfX3Ly/veL+eh8w/w6qq6vuPx+vt8oPUzXg33t+Cif3vone/uVfWrMfZ5OvALYKvRK6pqWbsVYy/gD4A/46EZBiPWAX5ZVbt2OIfxapckSZIkCZj8mQjbJNm9LR8AnAvMTrJ9a3sjcEFbvonerQAAr+7r405g477359G7BYIk6ybZBLgQ2Lc9c2Emvan+F61CnWcBb0+S1u8zV2HfEZcAL0qyebvN4gAeOrcVOZteAEA77q7t527Ay+jdMnFYkif179RmGcyqqv8E3kXv9grou1ZVdQdwY5L92z5J8oyV1DP6WkuSJEmS9KDJDhGuBd6cZDGwGfAxes8uODnJEnqzAT7dtn0/8PEkF9H7a/2IbwD7jTxYEXgn8OK2/yJg56q6HDgOWEjvl/ljq+qKVajzA8CjgMVJlrb3q6SqbgH+mt4tCFcBl1fV11ey2zuAue3hiNcAh6T3oMnPAgdV1U/pPRPhcyMBR7MxcHq7rhcAIw9B/DLwl+3hkNsBbwDekuQqerdyrPCBkc1iYFl7AKQPVpQkSZIkPUwemmU/wR33vlnh9Kp62qQcQAM1c+s5tfOhRw+6DEmSpE4Wzp836BIkaUpJsqiq5o5un+yZCJIkSZIkaZqYtAcrVtVNwJSdhZDkd+l9ZWO/G6tqv0HUI0mSJEnSoE32tzNMWVV1Fr0HLkqSJEmSJLydQZIkSZIkdWSIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUyYxBF6CpacctN2Hh/HmDLkOSJEmStAY5E0GSJEmSJHViiCBJkiRJkjoxRJAkSZIkSZ0YIkiSJEmSpE4MESRJkiRJUieGCJIkSZIkqRNDBEmSJEmS1IkhgiRJkiRJ6sQQQZIkSZIkdWKIIEmSJEmSOjFEkCRJkiRJnRgiSJIkSZKkTgwRJEmSJElSJ4YIkiRJkiSpE0MESZIkSZLUiSGCJEmSJEnqxBBBkiRJkiR1YoggSZIkSZI6MUSQJEmSJEmdpKoGXYOmoCR3AtcPug49aHPg1kEXoQc5HsPF8RgujsdwcTyGi+MxXByP4eJ4rHnbVtVjRzfOGEQlmhaur6q5gy5CPUkuczyGh+MxXByP4eJ4DBfHY7g4HsPF8Rgujsfw8HYGSZIkSZLUiSGCJEmSJEnqxBBBq+uYQRegh3E8hovjMVwcj+HieAwXx2O4OB7DxfEYLo7HkPDBipIkSZIkqRNnIkiSJEmSpE4MEdYiSX4vyfVJvp/k8DHWJ8kn2vrFSZ61sn2TbJbknCQ3tJ+/1bfur9v21yf53b72ZydZ0tZ9Ikkm87yH1TCMR5INk5yR5LokVyf50GSf97AahvEYdbzTkiydjHOdCoZlPJI8OskxSb7X/jt59WSe97AaovE4oH1+LE5yZpLNJ/O8h9WaHI8kj0nyrSR3JfnkqOP4ec5wjEf8PH/QMIzHqOP5eT4E4xE/zydWVflaC17AusAPgCcDjwauAnYatc3vA/8FBHgucMnK9gX+ATi8LR8OfLgt79S2Ww94Utt/3bZuIbB7O85/AS8b9PVZW8cD2BB4cdvm0cBFjsdg//to618FfAlYOuhrs7aPB/B+4O/a8jrA5oO+PmvreND7Wuqfj4xB2//IQV+ftWA8ZgJ7AIcAnxx1HD/Ph2Q88PN8qMaj71h+ng/JeODn+YS+nImw9tgN+H5V/bCqfg18Gdhn1Db7AF+onu8CmybZciX77gMc35aPB/bta/9yVd1XVTcC3wd2a/1tUlXfqd5/xV/o22dtMhTjUVX3VNW3AFpflwNPmITzHXZDMR4ASTYC3g383SSc51QxNOMBHAT8PUBVLa+qWyf4XKeCYRmPtNfM9hfvTYCfTvzpDr01Oh5VdXdVXQzc238AP88fNBTj4ef5g4ZiPMDP82ZoxgM/zyeUIcLaY2vg5r73P25tXbYZb98tquoWgPbzcR36+vFK6lgbDMt4PCjJpsDewHmrdirTwjCNxweAfwTuWZ0TmSaGYjzafxMAH0hyeZKTk2yxWmc0tQ3FeFTV/cCfAkvohQc7Af+2eqc0pa3p8RivDj/Ph2c8HuTn+dCMh5/nQzIefp5PPEOEtcdY9ymO/mqOFW3TZd+ux1udvqajYRmP3spkBvDvwCeq6ocr6Ws6GorxSLIrsH1VnbqS/ae7oRgPetPnnwAsqKpnAd8BPrqSvqajoRiPJI+iFyI8E9gKWAz89Ur6mo7W9Hg8kjrWBsMyHr0D+Xk+FOPh5/mDhmI88PN8whkirD1+DDyx7/0T+M1poCvaZrx9f9amHI1Mbfx5h76eMEb72mZYxmPEMcANVfXPq3oi08SwjMfuwLOT3ARcDMxJcv5qndHUNizj8Qt6f0Ea+UfgycCzWPsMy3jsClBVP2jT578CPG+1zmhqW9PjMV4dfp4Pz3iM8PN8OMbDz/OeYRkPP88nmCHC2uNS4ClJnpTk0cAfAKeN2uY04E3tKanPBW5vU4TG2/c04M1t+c3A1/va/yDJekmeBDwFWNj6uzPJc9s9rW/q22dtMhTjAZDk74BZwLsm4TyniqEYj6r6VFVtVVWz6T0Y6HtVtedknPCQG5bxKOAbwJ5tu72Aayb4XKeCoRgP4CfATkke27b7HeDaiT7ZKWBNj8eY/Dx/0FCMB/h53gzFePh5/qBhGQ8/zydaDcHTHX2tmRe9p59+j96TTue3tkOAQ9pygH9t65cAc8fbt7U/ht49dze0n5v1rZvftr+evicEA3OBpW3dJ4EM+tqsreNBL9Utev8Qv7K9/njQ12ZtHY9R9cxmLX2a8zCNB7AtcCG9qfPnAdsM+tqs5eNxSPv/q8X0/kH4mEFfm7VkPG4CbgPuovfXwZEnpPt5PiTjgZ/nQzUeo+qZjZ/nAx8P/Dyf0FfaRZUkSZIkSRqXtzNIkiRJkqRODBEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSerEEEGSJE1JSe5aw8ebneT1a/KYkiQNG0MESZKklUgyg973vRsiSJLWajMGXYAkSdIjkWRP4P3Az4Bdga8CS4B3AhsA+1bVD5IcB9wL7AxsAby7qk5Psj7wKWAusKy1fyvJgcDLgfWBmcCGwI5JrgSOB04FvtjWAfxZVX271XMkcCvwNGAR8IdVVUl+G/h42+c+YC/gHuBDwJ7AesC/VtVnJvIaSZI0UQwRJEnSdPAMYEfgNuCHwLFVtVuSdwJvB97VtpsNvAjYDvhWku2BtwFU1dOT7ACcnWRO2353YJequq2FA4dV1SsAkmwI/E5V3ZvkKcC/0wsiAJ5JL6z4KbAAeH6ShcBJwOuq6tIkmwC/At4C3F5Vv51kPWBBkrOr6sYJv0qSJD1ChgiSJGk6uLSqbgFI8gPg7Na+BHhx33ZfqarlwA1JfgjsAOwB/AtAVV2X5EfASIhwTlXdtoJjPgr4ZJJdgQf69gFYWFU/bvVcSS+8uB24paoubce6o61/KbBLkte0fWcBTwEMESRJQ8cQQZIkTQf39S0v73u/nIf/e6dG7VdAxun37nHW/Tm9WyieQe85U/euoJ4HWg0Z4/i09rdX1VnjHEuSpKHggxUlSdLaZP8k6yTZDngycD1wIfAGgHYbwzatfbQ7gY373s+iN7NgOfBGYN2VHPs6YKv2XASSbNwe2HgW8KdJHjVSQ5KZ4/QjSdLAOBNBkiStTa4HLqD3YMVD2vMMjgY+nWQJvQcrHlhV9yW/MUFhMbAsyVXAccDRwH8k2R/4FuPPWqCqfp3kdcC/JNmA3vMQ5gHH0rvd4fL0Dvq/wL4TcK6SJE24VI01q06SJGl6ad/OcHpVnTLoWiRJmqq8nUGSJEmSJHXiTARJkiRJktSJMxEkSZIkSVInhgiSJEmSJKkTQwRJkiRJktSJIYIkSZIkSerEEEGSJEmSJHViiCBJkiRJkjr5/ysPYEe4ijDfAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf = RandomForestClassifier(\n",
    "    n_estimators=500, \n",
    "    max_depth=2, \n",
    "    n_jobs=8, \n",
    "    random_state=42\n",
    ")\n",
    "rf.fit(X, y)\n",
    "fi = permutation_importance(\n",
    "    estimator=rf, \n",
    "    X=X, \n",
    "    y=y, \n",
    "    n_repeats=10, \n",
    "    n_jobs=8, \n",
    "    random_state=42\n",
    ").importances_mean\n",
    "\n",
    "figure(figsize=(16,8))\n",
    "barh(\n",
    "    y=range(10, 0, -1), \n",
    "    width=sorted(fi, reverse=True)[:10], \n",
    "    alpha=0.9\n",
    ")\n",
    "ylabel(\"Feature\")\n",
    "yticks(\n",
    "    range(10, 0, -1), \n",
    "    array_columns[fi.argsort()[::-1][:10]]\n",
    ")\n",
    "xlabel(\"Importance\")\n",
    "title(\"Original features importance\")\n",
    "show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "dim_layer_input = X.shape[1]\n",
    "dim_layer_1 = max((int(3*dim_layer_input/4), 1))\n",
    "dim_layer_2 = max((int(dim_layer_input/2), 1))\n",
    "autoencoder, encoder = build_autoencoder(\n",
    "    dim_input=dim_layer_input, \n",
    "    dim_layer_1=dim_layer_1, \n",
    "    dim_layer_2=dim_layer_2\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "113/113 [==============================] - 1s 4ms/step - loss: 0.1154 - val_loss: 0.0908\n",
      "Epoch 2/100\n",
      "113/113 [==============================] - 0s 4ms/step - loss: 0.0401 - val_loss: 0.0658\n",
      "Epoch 3/100\n",
      "113/113 [==============================] - 0s 4ms/step - loss: 0.0260 - val_loss: 0.0584\n",
      "Epoch 4/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0220 - val_loss: 0.0564\n",
      "Epoch 5/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0207 - val_loss: 0.0538\n",
      "Epoch 6/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0198 - val_loss: 0.0505\n",
      "Epoch 7/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0193 - val_loss: 0.0497\n",
      "Epoch 8/100\n",
      "113/113 [==============================] - 0s 4ms/step - loss: 0.0191 - val_loss: 0.0496\n",
      "Epoch 9/100\n",
      "113/113 [==============================] - 0s 4ms/step - loss: 0.0190 - val_loss: 0.0492\n",
      "Epoch 10/100\n",
      "113/113 [==============================] - 0s 4ms/step - loss: 0.0190 - val_loss: 0.0487\n",
      "Epoch 11/100\n",
      "113/113 [==============================] - 0s 4ms/step - loss: 0.0189 - val_loss: 0.0486\n",
      "Epoch 12/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0188 - val_loss: 0.0481\n",
      "Epoch 13/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0188 - val_loss: 0.0478\n",
      "Epoch 14/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0188 - val_loss: 0.0477\n",
      "Epoch 15/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0187 - val_loss: 0.0476\n",
      "Epoch 16/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0187 - val_loss: 0.0476\n",
      "Epoch 17/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0187 - val_loss: 0.0476\n",
      "Epoch 18/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0187 - val_loss: 0.0475\n",
      "Epoch 19/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0187 - val_loss: 0.0473\n",
      "Epoch 20/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0187 - val_loss: 0.0474\n",
      "Epoch 21/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0186 - val_loss: 0.0473\n",
      "Epoch 22/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0186 - val_loss: 0.0474\n",
      "Epoch 23/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0186 - val_loss: 0.0473\n",
      "Epoch 24/100\n",
      "113/113 [==============================] - 0s 3ms/step - loss: 0.0186 - val_loss: 0.0472\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7f1610687ca0>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "autoencoder.fit(\n",
    "    X, X,\n",
    "    epochs=100,\n",
    "    batch_size=256,\n",
    "    shuffle=True,\n",
    "    validation_split=0.3,\n",
    "    callbacks=callbacks\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoded = array(encoder(X))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = RandomForestClassifier(\n",
    "    n_estimators=500, \n",
    "    max_depth=2, \n",
    "    n_jobs=8, \n",
    "    random_state=42\n",
    ")\n",
    "dict_performance = cross_validate(\n",
    "    estimator=rf, \n",
    "    X=encoded, y=y, \n",
    "    cv=10, \n",
    "    n_jobs=4,\n",
    "    return_train_score=True,\n",
    "    scoring=[\n",
    "        \"balanced_accuracy\", \n",
    "        \"f1_weighted\", \n",
    "        \"roc_auc\", \n",
    "        \"average_precision\"\n",
    "    ]\n",
    ")\n",
    "\n",
    "df_performance[\"ENCODED\"] = [\n",
    "    mean(dict_performance[k]) \\\n",
    "    for k in dict_performance.keys()\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>ORIGINAL</th>\n",
       "      <th>ENCODED</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>fit_time</th>\n",
       "      <td>8.061133</td>\n",
       "      <td>26.388137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>score_time</th>\n",
       "      <td>0.830988</td>\n",
       "      <td>1.053183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_balanced_accuracy</th>\n",
       "      <td>0.505833</td>\n",
       "      <td>0.503019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_balanced_accuracy</th>\n",
       "      <td>0.517641</td>\n",
       "      <td>0.501363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_f1_weighted</th>\n",
       "      <td>0.809805</td>\n",
       "      <td>0.835511</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_f1_weighted</th>\n",
       "      <td>0.842227</td>\n",
       "      <td>0.835109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_roc_auc</th>\n",
       "      <td>0.563038</td>\n",
       "      <td>0.722752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_roc_auc</th>\n",
       "      <td>0.801313</td>\n",
       "      <td>0.762064</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_average_precision</th>\n",
       "      <td>0.414990</td>\n",
       "      <td>0.525588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_average_precision</th>\n",
       "      <td>0.463692</td>\n",
       "      <td>0.383266</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         ORIGINAL    ENCODED\n",
       "fit_time                 8.061133  26.388137\n",
       "score_time               0.830988   1.053183\n",
       "test_balanced_accuracy   0.505833   0.503019\n",
       "train_balanced_accuracy  0.517641   0.501363\n",
       "test_f1_weighted         0.809805   0.835511\n",
       "train_f1_weighted        0.842227   0.835109\n",
       "test_roc_auc             0.563038   0.722752\n",
       "train_roc_auc            0.801313   0.762064\n",
       "test_average_precision   0.414990   0.525588\n",
       "train_average_precision  0.463692   0.383266"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['age', 'campaign', 'pdays', 'previous', 'emp.var.rate',\n",
       "       'cons.price.idx', 'cons.conf.idx', 'euribor3m', 'nr.employed',\n",
       "       'job_admin.', 'job_blue-collar', 'job_entrepreneur', 'job_housemaid',\n",
       "       'job_management', 'job_retired', 'job_self-employed', 'job_services',\n",
       "       'job_student', 'job_technician', 'job_unemployed', 'job_unknown',\n",
       "       'marital_divorced', 'marital_married', 'marital_single',\n",
       "       'marital_unknown', 'education_basic.4y', 'education_basic.6y',\n",
       "       'education_basic.9y', 'education_high.school', 'education_illiterate',\n",
       "       'education_professional.course', 'education_university.degree',\n",
       "       'education_unknown', 'default_no', 'default_unknown', 'default_yes',\n",
       "       'housing_no', 'housing_unknown', 'housing_yes', 'loan_no',\n",
       "       'loan_unknown', 'loan_yes', 'contact_cellular', 'contact_telephone',\n",
       "       'month_apr', 'month_aug', 'month_dec', 'month_jul', 'month_jun',\n",
       "       'month_mar', 'month_may', 'month_nov', 'month_oct', 'month_sep',\n",
       "       'day_of_week_fri', 'day_of_week_mon', 'day_of_week_thu',\n",
       "       'day_of_week_tue', 'day_of_week_wed', 'poutcome_failure',\n",
       "       'poutcome_nonexistent', 'poutcome_success'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "array_columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "25 9 23 5 True\n"
     ]
    }
   ],
   "source": [
    "feat_pers = ['age', 'marital_divorced', 'marital_married', 'marital_single', 'marital_unknown', 'education_basic.4y', 'education_basic.6y', 'education_basic.9y', 'education_high.school', 'education_illiterate', 'education_professional.course', 'education_university.degree', 'education_unknown','job_admin.', 'job_blue-collar', 'job_entrepreneur', 'job_housemaid', 'job_management', 'job_retired', 'job_self-employed', 'job_services', 'job_student', 'job_technician', 'job_unemployed', 'job_unknown']\n",
    "feat_fina = ['default_no', 'default_unknown', 'default_yes', 'housing_no', 'housing_unknown', 'housing_yes', 'loan_no', 'loan_unknown', 'loan_yes']\n",
    "feat_camp = ['campaign', 'pdays', 'previous', 'contact_cellular', 'contact_telephone', 'month_apr', 'month_aug', 'month_dec', 'month_jul', 'month_jun', 'month_mar', 'month_may', 'month_nov', 'month_oct', 'month_sep', 'day_of_week_fri', 'day_of_week_mon', 'day_of_week_thu', 'day_of_week_tue', 'day_of_week_wed', 'poutcome_failure', 'poutcome_nonexistent', 'poutcome_success']\n",
    "feat_econ = ['emp.var.rate', 'cons.price.idx', 'cons.conf.idx', 'euribor3m', 'nr.employed']\n",
    "\n",
    "print(len(feat_pers), len(feat_fina), len(feat_camp), len(feat_econ), len(feat_pers) + len(feat_fina) + len(feat_camp) + len(feat_econ) == len(array_columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 4/4 [00:34<00:00,  8.51s/it]\n"
     ]
    }
   ],
   "source": [
    "feat_groups = [\n",
    "    feat_pers, \n",
    "    feat_fina, \n",
    "    feat_camp, \n",
    "    feat_econ\n",
    "]\n",
    "encoded = []\n",
    "\n",
    "for g in tqdm(feat_groups):\n",
    "    \n",
    "    dim_layer_input = len(g)\n",
    "    dim_layer_1 = max((int(3*dim_layer_input/4), 1))\n",
    "    dim_layer_2 = max((int(dim_layer_input / 2), 1))\n",
    "    \n",
    "    autoencoder, encoder = build_autoencoder(\n",
    "        dim_input=dim_layer_input, \n",
    "        dim_layer_1=dim_layer_1, \n",
    "        dim_layer_2=dim_layer_2\n",
    "    )\n",
    "    \n",
    "    X_tmp = X[:, array_columns.isin(g)]\n",
    "    autoencoder.fit(\n",
    "        X_tmp, X_tmp,\n",
    "        epochs=100,\n",
    "        batch_size=256,\n",
    "        shuffle=True,\n",
    "        validation_split=0.3,\n",
    "        callbacks=callbacks, \n",
    "        verbose=0\n",
    "    )\n",
    "    \n",
    "    encoded.append(array(encoder(X_tmp)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_encoded = hstack(encoded)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "array_encoded_features = array(\n",
    "    [\"pers_\"+str(j) for j in range(encoded[0].shape[1])] + \\\n",
    "    [\"fina_\"+str(j) for j in range(encoded[1].shape[1])] + \\\n",
    "    [\"camp_\"+str(j) for j in range(encoded[2].shape[1])] + \\\n",
    "    [\"econ_\"+str(j) for j in range(encoded[3].shape[1])]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "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>pers_0</th>\n",
       "      <th>pers_1</th>\n",
       "      <th>pers_2</th>\n",
       "      <th>pers_3</th>\n",
       "      <th>pers_4</th>\n",
       "      <th>pers_5</th>\n",
       "      <th>pers_6</th>\n",
       "      <th>pers_7</th>\n",
       "      <th>pers_8</th>\n",
       "      <th>pers_9</th>\n",
       "      <th>pers_10</th>\n",
       "      <th>pers_11</th>\n",
       "      <th>fina_0</th>\n",
       "      <th>fina_1</th>\n",
       "      <th>fina_2</th>\n",
       "      <th>fina_3</th>\n",
       "      <th>camp_0</th>\n",
       "      <th>camp_1</th>\n",
       "      <th>camp_2</th>\n",
       "      <th>camp_3</th>\n",
       "      <th>camp_4</th>\n",
       "      <th>camp_5</th>\n",
       "      <th>camp_6</th>\n",
       "      <th>camp_7</th>\n",
       "      <th>camp_8</th>\n",
       "      <th>camp_9</th>\n",
       "      <th>camp_10</th>\n",
       "      <th>econ_0</th>\n",
       "      <th>econ_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14271</th>\n",
       "      <td>1.766457</td>\n",
       "      <td>-1.858506</td>\n",
       "      <td>1.527646</td>\n",
       "      <td>0.967267</td>\n",
       "      <td>1.197702</td>\n",
       "      <td>1.685395</td>\n",
       "      <td>1.742300</td>\n",
       "      <td>1.521291</td>\n",
       "      <td>3.352483</td>\n",
       "      <td>3.295268</td>\n",
       "      <td>2.039039</td>\n",
       "      <td>0.819601</td>\n",
       "      <td>-1.159335</td>\n",
       "      <td>3.072608</td>\n",
       "      <td>2.940951</td>\n",
       "      <td>1.715309</td>\n",
       "      <td>0.756437</td>\n",
       "      <td>1.554016</td>\n",
       "      <td>0.125771</td>\n",
       "      <td>0.818307</td>\n",
       "      <td>-0.655496</td>\n",
       "      <td>-0.257394</td>\n",
       "      <td>2.263055</td>\n",
       "      <td>1.110901</td>\n",
       "      <td>1.666754</td>\n",
       "      <td>2.315678</td>\n",
       "      <td>0.594455</td>\n",
       "      <td>-0.073504</td>\n",
       "      <td>0.706128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38110</th>\n",
       "      <td>1.900608</td>\n",
       "      <td>-1.289555</td>\n",
       "      <td>0.934587</td>\n",
       "      <td>1.264449</td>\n",
       "      <td>0.917077</td>\n",
       "      <td>2.590348</td>\n",
       "      <td>1.985073</td>\n",
       "      <td>0.344954</td>\n",
       "      <td>3.171203</td>\n",
       "      <td>1.493695</td>\n",
       "      <td>1.479075</td>\n",
       "      <td>1.615082</td>\n",
       "      <td>-0.655238</td>\n",
       "      <td>3.922607</td>\n",
       "      <td>1.190354</td>\n",
       "      <td>3.957587</td>\n",
       "      <td>0.836479</td>\n",
       "      <td>1.537111</td>\n",
       "      <td>0.096256</td>\n",
       "      <td>0.534085</td>\n",
       "      <td>-0.718423</td>\n",
       "      <td>0.121315</td>\n",
       "      <td>1.322045</td>\n",
       "      <td>-0.155505</td>\n",
       "      <td>0.959476</td>\n",
       "      <td>1.752178</td>\n",
       "      <td>-0.003634</td>\n",
       "      <td>-0.032568</td>\n",
       "      <td>0.069207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9613</th>\n",
       "      <td>0.727200</td>\n",
       "      <td>-1.494521</td>\n",
       "      <td>0.445399</td>\n",
       "      <td>2.098301</td>\n",
       "      <td>1.099392</td>\n",
       "      <td>1.233449</td>\n",
       "      <td>2.386735</td>\n",
       "      <td>0.855103</td>\n",
       "      <td>2.531948</td>\n",
       "      <td>2.135375</td>\n",
       "      <td>1.562708</td>\n",
       "      <td>0.578137</td>\n",
       "      <td>1.579593</td>\n",
       "      <td>6.062159</td>\n",
       "      <td>-1.143997</td>\n",
       "      <td>0.049949</td>\n",
       "      <td>1.806336</td>\n",
       "      <td>0.627548</td>\n",
       "      <td>1.027021</td>\n",
       "      <td>1.162885</td>\n",
       "      <td>-0.387225</td>\n",
       "      <td>-0.350623</td>\n",
       "      <td>2.001506</td>\n",
       "      <td>1.404485</td>\n",
       "      <td>1.161019</td>\n",
       "      <td>2.686127</td>\n",
       "      <td>0.844868</td>\n",
       "      <td>-0.015803</td>\n",
       "      <td>0.810468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19167</th>\n",
       "      <td>3.136367</td>\n",
       "      <td>-1.855610</td>\n",
       "      <td>0.047800</td>\n",
       "      <td>3.380212</td>\n",
       "      <td>0.777369</td>\n",
       "      <td>1.076406</td>\n",
       "      <td>1.276556</td>\n",
       "      <td>1.325401</td>\n",
       "      <td>1.310340</td>\n",
       "      <td>2.068194</td>\n",
       "      <td>3.934725</td>\n",
       "      <td>0.668451</td>\n",
       "      <td>-0.655238</td>\n",
       "      <td>3.922607</td>\n",
       "      <td>1.190354</td>\n",
       "      <td>3.957587</td>\n",
       "      <td>0.930507</td>\n",
       "      <td>1.777684</td>\n",
       "      <td>-0.117286</td>\n",
       "      <td>0.521362</td>\n",
       "      <td>-0.757560</td>\n",
       "      <td>-0.009650</td>\n",
       "      <td>1.765950</td>\n",
       "      <td>-0.012735</td>\n",
       "      <td>0.811082</td>\n",
       "      <td>1.673729</td>\n",
       "      <td>0.311292</td>\n",
       "      <td>-0.076619</td>\n",
       "      <td>0.657558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11861</th>\n",
       "      <td>3.687668</td>\n",
       "      <td>-1.844795</td>\n",
       "      <td>0.373193</td>\n",
       "      <td>2.627980</td>\n",
       "      <td>0.973278</td>\n",
       "      <td>1.552355</td>\n",
       "      <td>1.002753</td>\n",
       "      <td>0.830468</td>\n",
       "      <td>1.814662</td>\n",
       "      <td>2.220145</td>\n",
       "      <td>4.481116</td>\n",
       "      <td>0.567920</td>\n",
       "      <td>-0.865288</td>\n",
       "      <td>2.007591</td>\n",
       "      <td>1.253906</td>\n",
       "      <td>2.867021</td>\n",
       "      <td>2.064615</td>\n",
       "      <td>1.125823</td>\n",
       "      <td>0.756587</td>\n",
       "      <td>1.146341</td>\n",
       "      <td>-0.359168</td>\n",
       "      <td>-0.276827</td>\n",
       "      <td>1.921229</td>\n",
       "      <td>0.782431</td>\n",
       "      <td>0.134812</td>\n",
       "      <td>1.864548</td>\n",
       "      <td>1.563612</td>\n",
       "      <td>-0.015927</td>\n",
       "      <td>0.810330</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         pers_0    pers_1    pers_2    pers_3    pers_4    pers_5    pers_6  \\\n",
       "14271  1.766457 -1.858506  1.527646  0.967267  1.197702  1.685395  1.742300   \n",
       "38110  1.900608 -1.289555  0.934587  1.264449  0.917077  2.590348  1.985073   \n",
       "9613   0.727200 -1.494521  0.445399  2.098301  1.099392  1.233449  2.386735   \n",
       "19167  3.136367 -1.855610  0.047800  3.380212  0.777369  1.076406  1.276556   \n",
       "11861  3.687668 -1.844795  0.373193  2.627980  0.973278  1.552355  1.002753   \n",
       "\n",
       "         pers_7    pers_8    pers_9   pers_10   pers_11    fina_0    fina_1  \\\n",
       "14271  1.521291  3.352483  3.295268  2.039039  0.819601 -1.159335  3.072608   \n",
       "38110  0.344954  3.171203  1.493695  1.479075  1.615082 -0.655238  3.922607   \n",
       "9613   0.855103  2.531948  2.135375  1.562708  0.578137  1.579593  6.062159   \n",
       "19167  1.325401  1.310340  2.068194  3.934725  0.668451 -0.655238  3.922607   \n",
       "11861  0.830468  1.814662  2.220145  4.481116  0.567920 -0.865288  2.007591   \n",
       "\n",
       "         fina_2    fina_3    camp_0    camp_1    camp_2    camp_3    camp_4  \\\n",
       "14271  2.940951  1.715309  0.756437  1.554016  0.125771  0.818307 -0.655496   \n",
       "38110  1.190354  3.957587  0.836479  1.537111  0.096256  0.534085 -0.718423   \n",
       "9613  -1.143997  0.049949  1.806336  0.627548  1.027021  1.162885 -0.387225   \n",
       "19167  1.190354  3.957587  0.930507  1.777684 -0.117286  0.521362 -0.757560   \n",
       "11861  1.253906  2.867021  2.064615  1.125823  0.756587  1.146341 -0.359168   \n",
       "\n",
       "         camp_5    camp_6    camp_7    camp_8    camp_9   camp_10    econ_0  \\\n",
       "14271 -0.257394  2.263055  1.110901  1.666754  2.315678  0.594455 -0.073504   \n",
       "38110  0.121315  1.322045 -0.155505  0.959476  1.752178 -0.003634 -0.032568   \n",
       "9613  -0.350623  2.001506  1.404485  1.161019  2.686127  0.844868 -0.015803   \n",
       "19167 -0.009650  1.765950 -0.012735  0.811082  1.673729  0.311292 -0.076619   \n",
       "11861 -0.276827  1.921229  0.782431  0.134812  1.864548  1.563612 -0.015927   \n",
       "\n",
       "         econ_1  \n",
       "14271  0.706128  \n",
       "38110  0.069207  \n",
       "9613   0.810468  \n",
       "19167  0.657558  \n",
       "11861  0.810330  "
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DataFrame(X_encoded, columns=array_encoded_features).sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(41188, 29)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_encoded.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(41188, 62)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = RandomForestClassifier(n_estimators=500, max_depth=2, n_jobs=8, random_state=42)\n",
    "\n",
    "dict_performance = cross_validate(\n",
    "    estimator=rf, \n",
    "    X=X_encoded, y=y, \n",
    "    cv=10, \n",
    "    n_jobs=4,\n",
    "    return_train_score=True,\n",
    "    scoring=[\n",
    "        \"balanced_accuracy\", \n",
    "        \"f1_weighted\", \n",
    "        \"roc_auc\", \n",
    "        \"average_precision\"\n",
    "    ]\n",
    ")\n",
    "\n",
    "df_performance[\"GROUP_ENCODED\"] = [mean(dict_performance[k]) for k in dict_performance.keys()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>ORIGINAL</th>\n",
       "      <th>ENCODED</th>\n",
       "      <th>GROUP_ENCODED</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>fit_time</th>\n",
       "      <td>8.061133</td>\n",
       "      <td>26.388137</td>\n",
       "      <td>14.218396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>score_time</th>\n",
       "      <td>0.830988</td>\n",
       "      <td>1.053183</td>\n",
       "      <td>0.853897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_balanced_accuracy</th>\n",
       "      <td>0.505833</td>\n",
       "      <td>0.503019</td>\n",
       "      <td>0.523597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_balanced_accuracy</th>\n",
       "      <td>0.517641</td>\n",
       "      <td>0.501363</td>\n",
       "      <td>0.528023</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_f1_weighted</th>\n",
       "      <td>0.809805</td>\n",
       "      <td>0.835511</td>\n",
       "      <td>0.838770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_f1_weighted</th>\n",
       "      <td>0.842227</td>\n",
       "      <td>0.835109</td>\n",
       "      <td>0.848629</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_roc_auc</th>\n",
       "      <td>0.563038</td>\n",
       "      <td>0.722752</td>\n",
       "      <td>0.493781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_roc_auc</th>\n",
       "      <td>0.801313</td>\n",
       "      <td>0.762064</td>\n",
       "      <td>0.791968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_average_precision</th>\n",
       "      <td>0.414990</td>\n",
       "      <td>0.525588</td>\n",
       "      <td>0.303416</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>train_average_precision</th>\n",
       "      <td>0.463692</td>\n",
       "      <td>0.383266</td>\n",
       "      <td>0.454403</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         ORIGINAL    ENCODED  GROUP_ENCODED\n",
       "fit_time                 8.061133  26.388137      14.218396\n",
       "score_time               0.830988   1.053183       0.853897\n",
       "test_balanced_accuracy   0.505833   0.503019       0.523597\n",
       "train_balanced_accuracy  0.517641   0.501363       0.528023\n",
       "test_f1_weighted         0.809805   0.835511       0.838770\n",
       "train_f1_weighted        0.842227   0.835109       0.848629\n",
       "test_roc_auc             0.563038   0.722752       0.493781\n",
       "train_roc_auc            0.801313   0.762064       0.791968\n",
       "test_average_precision   0.414990   0.525588       0.303416\n",
       "train_average_precision  0.463692   0.383266       0.454403"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf = RandomForestClassifier(\n",
    "    n_estimators=500, \n",
    "    max_depth=2, \n",
    "    n_jobs=8, \n",
    "    random_state=42\n",
    ")\n",
    "rf.fit(X_encoded, y)\n",
    "fi = permutation_importance(\n",
    "    estimator=rf, \n",
    "    X=X_encoded, \n",
    "    y=y, \n",
    "    n_repeats=10, \n",
    "    n_jobs=8, \n",
    "    random_state=42\n",
    ").importances_mean\n",
    "\n",
    "figure(figsize=(16,8))\n",
    "barh(\n",
    "    y=range(10, 0, -1), \n",
    "    width=sorted(fi, reverse=True)[:10], \n",
    "    alpha=0.9\n",
    ")\n",
    "ylabel(\"Feature\")\n",
    "yticks(\n",
    "    range(10, 0, -1), \n",
    "    array_encoded_features[fi.argsort()[::-1][:10]]\n",
    ")\n",
    "xlabel(\"Importance\")\n",
    "title(\"Group-encoded features importance\")\n",
    "show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}