Stage2_MachineLearningModel.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "e14620c8",
   "metadata": {},
   "source": [
    "# Churn Prediction Using Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1000f97c",
   "metadata": {},
   "source": [
    "In here, I am going to use two methods, using the actual samples and using resampling method to make each class have equal samples. The expectation when using resampling method is to avoid bias because when testing the models with actual samples, they tend to predict all samples to not churn (because the data contain more `0` class)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "00c89be5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# importing libraries\n",
    "# main libraries\n",
    "import pandas as pd\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "# model building libraries\n",
    "from sklearn.feature_selection import RFE, RFECV\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier\n",
    "from sklearn.pipeline import Pipeline\n",
    "%matplotlib inline\n",
    "plt.rcParams['figure.figsize']=(10,6)\n",
    "\n",
    "from warnings import filterwarnings\n",
    "filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b349634d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialise categorical columns\n",
    "cat_cols = ['HasCrCard', 'IsActiveMember', 'Complain', 'Card Type', 'Female', 'Male', \n",
    "        'France','Germany','Spain','zero_balance','high_balance_low_salary','high_salary_low_balance']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "db530069",
   "metadata": {
    "scrolled": true
   },
   "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>CustomerId</th>\n",
       "      <th>Surname</th>\n",
       "      <th>CreditScore</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>HasCrCard</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>Exited</th>\n",
       "      <th>Complain</th>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <th>Card Type</th>\n",
       "      <th>Point Earned</th>\n",
       "      <th>Female</th>\n",
       "      <th>Male</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Spain</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>15634602</td>\n",
       "      <td>Hargrave</td>\n",
       "      <td>619</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>464</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15647311</td>\n",
       "      <td>Hill</td>\n",
       "      <td>608</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>456</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15619304</td>\n",
       "      <td>Onio</td>\n",
       "      <td>502</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>377</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>15701354</td>\n",
       "      <td>Boni</td>\n",
       "      <td>699</td>\n",
       "      <td>39</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>93826.63</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>GOLD</td>\n",
       "      <td>350</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15737888</td>\n",
       "      <td>Mitchell</td>\n",
       "      <td>850</td>\n",
       "      <td>43</td>\n",
       "      <td>2</td>\n",
       "      <td>125510.82</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>79084.10</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>GOLD</td>\n",
       "      <td>425</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustomerId   Surname  CreditScore  Age  Tenure    Balance  NumOfProducts  \\\n",
       "0    15634602  Hargrave          619   42       2       0.00              1   \n",
       "1    15647311      Hill          608   41       1   83807.86              1   \n",
       "2    15619304      Onio          502   42       8  159660.80              3   \n",
       "3    15701354      Boni          699   39       1       0.00              2   \n",
       "4    15737888  Mitchell          850   43       2  125510.82              1   \n",
       "\n",
       "   HasCrCard  IsActiveMember  EstimatedSalary  Exited  Complain  \\\n",
       "0          1               1        101348.88       1         1   \n",
       "1          0               1        112542.58       0         1   \n",
       "2          1               0        113931.57       1         1   \n",
       "3          0               0         93826.63       0         0   \n",
       "4          1               1         79084.10       0         0   \n",
       "\n",
       "   Satisfaction Score Card Type  Point Earned  Female  Male  France  Germany  \\\n",
       "0                   2   DIAMOND           464       1     0       1        0   \n",
       "1                   3   DIAMOND           456       1     0       0        0   \n",
       "2                   3   DIAMOND           377       1     0       1        0   \n",
       "3                   5      GOLD           350       1     0       1        0   \n",
       "4                   5      GOLD           425       1     0       0        0   \n",
       "\n",
       "   Spain  \n",
       "0      0  \n",
       "1      1  \n",
       "2      0  \n",
       "3      0  \n",
       "4      1  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('PrepChurnData.csv')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8de16490",
   "metadata": {},
   "source": [
    "# Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "13e7e5ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import normalize, LabelEncoder, StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "e0b73644",
   "metadata": {
    "scrolled": true
   },
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>CustomerId</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.569094e+07</td>\n",
       "      <td>71936.186123</td>\n",
       "      <td>15565701.00</td>\n",
       "      <td>15628528.25</td>\n",
       "      <td>1.569074e+07</td>\n",
       "      <td>1.575323e+07</td>\n",
       "      <td>15815690.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CreditScore</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>6.505288e+02</td>\n",
       "      <td>96.653299</td>\n",
       "      <td>350.00</td>\n",
       "      <td>584.00</td>\n",
       "      <td>6.520000e+02</td>\n",
       "      <td>7.180000e+02</td>\n",
       "      <td>850.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Age</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>3.892180e+01</td>\n",
       "      <td>10.487806</td>\n",
       "      <td>18.00</td>\n",
       "      <td>32.00</td>\n",
       "      <td>3.700000e+01</td>\n",
       "      <td>4.400000e+01</td>\n",
       "      <td>92.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Tenure</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.012800e+00</td>\n",
       "      <td>2.892174</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>5.000000e+00</td>\n",
       "      <td>7.000000e+00</td>\n",
       "      <td>10.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Balance</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>7.648589e+04</td>\n",
       "      <td>62397.405202</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>9.719854e+04</td>\n",
       "      <td>1.276442e+05</td>\n",
       "      <td>250898.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NumOfProducts</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.530200e+00</td>\n",
       "      <td>0.581654</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>2.000000e+00</td>\n",
       "      <td>4.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HasCrCard</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>7.055000e-01</td>\n",
       "      <td>0.455840</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IsActiveMember</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.151000e-01</td>\n",
       "      <td>0.499797</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.000902e+05</td>\n",
       "      <td>57510.492818</td>\n",
       "      <td>11.58</td>\n",
       "      <td>51002.11</td>\n",
       "      <td>1.001939e+05</td>\n",
       "      <td>1.493882e+05</td>\n",
       "      <td>199992.48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Exited</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.038000e-01</td>\n",
       "      <td>0.402842</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Complain</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.044000e-01</td>\n",
       "      <td>0.403283</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>3.013800e+00</td>\n",
       "      <td>1.405919</td>\n",
       "      <td>1.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>3.000000e+00</td>\n",
       "      <td>4.000000e+00</td>\n",
       "      <td>5.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Point Earned</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>6.065151e+02</td>\n",
       "      <td>225.924839</td>\n",
       "      <td>119.00</td>\n",
       "      <td>410.00</td>\n",
       "      <td>6.050000e+02</td>\n",
       "      <td>8.010000e+02</td>\n",
       "      <td>1000.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Female</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>4.543000e-01</td>\n",
       "      <td>0.497932</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Male</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.457000e-01</td>\n",
       "      <td>0.497932</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.014000e-01</td>\n",
       "      <td>0.500023</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Germany</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.509000e-01</td>\n",
       "      <td>0.433553</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Spain</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.477000e-01</td>\n",
       "      <td>0.431698</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      count          mean           std          min  \\\n",
       "CustomerId          10000.0  1.569094e+07  71936.186123  15565701.00   \n",
       "CreditScore         10000.0  6.505288e+02     96.653299       350.00   \n",
       "Age                 10000.0  3.892180e+01     10.487806        18.00   \n",
       "Tenure              10000.0  5.012800e+00      2.892174         0.00   \n",
       "Balance             10000.0  7.648589e+04  62397.405202         0.00   \n",
       "NumOfProducts       10000.0  1.530200e+00      0.581654         1.00   \n",
       "HasCrCard           10000.0  7.055000e-01      0.455840         0.00   \n",
       "IsActiveMember      10000.0  5.151000e-01      0.499797         0.00   \n",
       "EstimatedSalary     10000.0  1.000902e+05  57510.492818        11.58   \n",
       "Exited              10000.0  2.038000e-01      0.402842         0.00   \n",
       "Complain            10000.0  2.044000e-01      0.403283         0.00   \n",
       "Satisfaction Score  10000.0  3.013800e+00      1.405919         1.00   \n",
       "Point Earned        10000.0  6.065151e+02    225.924839       119.00   \n",
       "Female              10000.0  4.543000e-01      0.497932         0.00   \n",
       "Male                10000.0  5.457000e-01      0.497932         0.00   \n",
       "France              10000.0  5.014000e-01      0.500023         0.00   \n",
       "Germany             10000.0  2.509000e-01      0.433553         0.00   \n",
       "Spain               10000.0  2.477000e-01      0.431698         0.00   \n",
       "\n",
       "                            25%           50%           75%          max  \n",
       "CustomerId          15628528.25  1.569074e+07  1.575323e+07  15815690.00  \n",
       "CreditScore              584.00  6.520000e+02  7.180000e+02       850.00  \n",
       "Age                       32.00  3.700000e+01  4.400000e+01        92.00  \n",
       "Tenure                     3.00  5.000000e+00  7.000000e+00        10.00  \n",
       "Balance                    0.00  9.719854e+04  1.276442e+05    250898.09  \n",
       "NumOfProducts              1.00  1.000000e+00  2.000000e+00         4.00  \n",
       "HasCrCard                  0.00  1.000000e+00  1.000000e+00         1.00  \n",
       "IsActiveMember             0.00  1.000000e+00  1.000000e+00         1.00  \n",
       "EstimatedSalary        51002.11  1.001939e+05  1.493882e+05    199992.48  \n",
       "Exited                     0.00  0.000000e+00  0.000000e+00         1.00  \n",
       "Complain                   0.00  0.000000e+00  0.000000e+00         1.00  \n",
       "Satisfaction Score         2.00  3.000000e+00  4.000000e+00         5.00  \n",
       "Point Earned             410.00  6.050000e+02  8.010000e+02      1000.00  \n",
       "Female                     0.00  0.000000e+00  1.000000e+00         1.00  \n",
       "Male                       0.00  1.000000e+00  1.000000e+00         1.00  \n",
       "France                     0.00  1.000000e+00  1.000000e+00         1.00  \n",
       "Germany                    0.00  0.000000e+00  1.000000e+00         1.00  \n",
       "Spain                      0.00  0.000000e+00  0.000000e+00         1.00  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.describe().T"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a52cbabe",
   "metadata": {},
   "source": [
    "We want to transform feature(s) that have high values and normalise them for our models. This will make our models interpret all features with similar weights."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "56c26687",
   "metadata": {},
   "outputs": [],
   "source": [
    "def engineer(df):\n",
    "    data = df.copy()\n",
    "    high_values = ['Tenure','CreditScore', 'Balance', 'EstimatedSalary', 'Point Earned']\n",
    "    \n",
    "    data['zero_balance'] = data['Balance']==0\n",
    "    data['female_male_ratio'] = data.Female.sum() / data.Male.sum()\n",
    "    \n",
    "    for col in high_values:\n",
    "        data[col] = normalize(np.array(data[col]).reshape(1,-1)).reshape(-1,1)\n",
    "\n",
    "    data['tenure_balance_ratio'] = np.where(data.zero_balance==False, data.Tenure/data.Balance, 0)\n",
    "    data['salary_balance_ratio'] = np.where(data.zero_balance==False, data.EstimatedSalary/data.Balance, 0)\n",
    "    data['high_balance_low_salary'] = np.where(data.salary_balance_ratio==0, 1, 0)\n",
    "    data['high_salary_low_balance'] = np.where(data.salary_balance_ratio>=1, 1, 0)\n",
    "    enc = LabelEncoder()\n",
    "    data['Card Type'] = enc.fit_transform(data['Card Type'])\n",
    "#     data['complain_score'] = np.where(data['Complain']==1, \n",
    "#                                       data.loc[data['Complain']==1,'Satisfaction Score'].mean(),\n",
    "#                                       0)\n",
    "#     data['complain_score'] = np.where(data['complain_score']==0, \n",
    "#                                       data.loc[data['Complain']==0,'Satisfaction Score'].mean(),\n",
    "#                                       data['complain_score'])\n",
    "    \n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "28f44aee",
   "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>CustomerId</th>\n",
       "      <th>Surname</th>\n",
       "      <th>CreditScore</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>HasCrCard</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>...</th>\n",
       "      <th>Male</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Spain</th>\n",
       "      <th>zero_balance</th>\n",
       "      <th>female_male_ratio</th>\n",
       "      <th>tenure_balance_ratio</th>\n",
       "      <th>salary_balance_ratio</th>\n",
       "      <th>high_balance_low_salary</th>\n",
       "      <th>high_salary_low_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>15634602</td>\n",
       "      <td>Hargrave</td>\n",
       "      <td>0.009412</td>\n",
       "      <td>42</td>\n",
       "      <td>0.003456</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.008780</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>True</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15647311</td>\n",
       "      <td>Hill</td>\n",
       "      <td>0.009245</td>\n",
       "      <td>41</td>\n",
       "      <td>0.001728</td>\n",
       "      <td>0.008491</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.009749</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.203514</td>\n",
       "      <td>1.148273</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15619304</td>\n",
       "      <td>Onio</td>\n",
       "      <td>0.007633</td>\n",
       "      <td>42</td>\n",
       "      <td>0.013824</td>\n",
       "      <td>0.016175</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.009870</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.854615</td>\n",
       "      <td>0.610181</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>15701354</td>\n",
       "      <td>Boni</td>\n",
       "      <td>0.010628</td>\n",
       "      <td>39</td>\n",
       "      <td>0.001728</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.008128</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>True</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15737888</td>\n",
       "      <td>Mitchell</td>\n",
       "      <td>0.012924</td>\n",
       "      <td>43</td>\n",
       "      <td>0.003456</td>\n",
       "      <td>0.012715</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.006851</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.271786</td>\n",
       "      <td>0.538792</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustomerId   Surname  CreditScore  Age    Tenure   Balance  NumOfProducts  \\\n",
       "0    15634602  Hargrave     0.009412   42  0.003456  0.000000              1   \n",
       "1    15647311      Hill     0.009245   41  0.001728  0.008491              1   \n",
       "2    15619304      Onio     0.007633   42  0.013824  0.016175              3   \n",
       "3    15701354      Boni     0.010628   39  0.001728  0.000000              2   \n",
       "4    15737888  Mitchell     0.012924   43  0.003456  0.012715              1   \n",
       "\n",
       "   HasCrCard  IsActiveMember  EstimatedSalary  ...  Male  France  Germany  \\\n",
       "0          1               1         0.008780  ...     0       1        0   \n",
       "1          0               1         0.009749  ...     0       0        0   \n",
       "2          1               0         0.009870  ...     0       1        0   \n",
       "3          0               0         0.008128  ...     0       1        0   \n",
       "4          1               1         0.006851  ...     0       0        0   \n",
       "\n",
       "   Spain  zero_balance  female_male_ratio  tenure_balance_ratio  \\\n",
       "0      0          True           0.832509              0.000000   \n",
       "1      1         False           0.832509              0.203514   \n",
       "2      0         False           0.832509              0.854615   \n",
       "3      0          True           0.832509              0.000000   \n",
       "4      1         False           0.832509              0.271786   \n",
       "\n",
       "   salary_balance_ratio  high_balance_low_salary  high_salary_low_balance  \n",
       "0              0.000000                        1                        0  \n",
       "1              1.148273                        0                        1  \n",
       "2              0.610181                        0                        0  \n",
       "3              0.000000                        1                        0  \n",
       "4              0.538792                        0                        0  \n",
       "\n",
       "[5 rows x 26 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eng_data = engineer(data)\n",
    "eng_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f5965deb",
   "metadata": {
    "scrolled": true
   },
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>CustomerId</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.569094e+07</td>\n",
       "      <td>7.193619e+04</td>\n",
       "      <td>1.556570e+07</td>\n",
       "      <td>1.562853e+07</td>\n",
       "      <td>1.569074e+07</td>\n",
       "      <td>1.575323e+07</td>\n",
       "      <td>1.581569e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CreditScore</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>9.891430e-03</td>\n",
       "      <td>1.469634e-03</td>\n",
       "      <td>5.321825e-03</td>\n",
       "      <td>8.879845e-03</td>\n",
       "      <td>9.913800e-03</td>\n",
       "      <td>1.091734e-02</td>\n",
       "      <td>1.292443e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Age</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>3.892180e+01</td>\n",
       "      <td>1.048781e+01</td>\n",
       "      <td>1.800000e+01</td>\n",
       "      <td>3.200000e+01</td>\n",
       "      <td>3.700000e+01</td>\n",
       "      <td>4.400000e+01</td>\n",
       "      <td>9.200000e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Tenure</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>8.661834e-03</td>\n",
       "      <td>4.997513e-03</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>5.183830e-03</td>\n",
       "      <td>8.639716e-03</td>\n",
       "      <td>1.209560e-02</td>\n",
       "      <td>1.727943e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Balance</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>7.748754e-03</td>\n",
       "      <td>6.321456e-03</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>9.847145e-03</td>\n",
       "      <td>1.293159e-02</td>\n",
       "      <td>2.541838e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NumOfProducts</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.530200e+00</td>\n",
       "      <td>5.816544e-01</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>2.000000e+00</td>\n",
       "      <td>4.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HasCrCard</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>7.055000e-01</td>\n",
       "      <td>4.558405e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IsActiveMember</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.151000e-01</td>\n",
       "      <td>4.997969e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>8.670720e-03</td>\n",
       "      <td>4.982078e-03</td>\n",
       "      <td>1.003164e-06</td>\n",
       "      <td>4.418263e-03</td>\n",
       "      <td>8.679701e-03</td>\n",
       "      <td>1.294136e-02</td>\n",
       "      <td>1.732515e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Exited</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.038000e-01</td>\n",
       "      <td>4.028421e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Complain</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.044000e-01</td>\n",
       "      <td>4.032827e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>3.013800e+00</td>\n",
       "      <td>1.405919e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>2.000000e+00</td>\n",
       "      <td>3.000000e+00</td>\n",
       "      <td>4.000000e+00</td>\n",
       "      <td>5.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Card Type</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.498000e+00</td>\n",
       "      <td>1.118356e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>2.000000e+00</td>\n",
       "      <td>3.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Point Earned</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>9.371039e-03</td>\n",
       "      <td>3.490681e-03</td>\n",
       "      <td>1.838625e-03</td>\n",
       "      <td>6.334757e-03</td>\n",
       "      <td>9.347630e-03</td>\n",
       "      <td>1.237595e-02</td>\n",
       "      <td>1.545063e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Female</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>4.543000e-01</td>\n",
       "      <td>4.979320e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Male</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.457000e-01</td>\n",
       "      <td>4.979320e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.014000e-01</td>\n",
       "      <td>5.000230e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Germany</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.509000e-01</td>\n",
       "      <td>4.335527e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Spain</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>2.477000e-01</td>\n",
       "      <td>4.316982e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>female_male_ratio</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>8.325087e-01</td>\n",
       "      <td>2.116191e-13</td>\n",
       "      <td>8.325087e-01</td>\n",
       "      <td>8.325087e-01</td>\n",
       "      <td>8.325087e-01</td>\n",
       "      <td>8.325087e-01</td>\n",
       "      <td>8.325087e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tenure_balance_ratio</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>4.892430e-01</td>\n",
       "      <td>5.755533e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>3.027190e-01</td>\n",
       "      <td>8.620636e-01</td>\n",
       "      <td>9.566742e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>salary_balance_ratio</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.007213e-01</td>\n",
       "      <td>7.014547e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>3.186064e-01</td>\n",
       "      <td>8.764975e-01</td>\n",
       "      <td>4.017499e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>high_balance_low_salary</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>3.617000e-01</td>\n",
       "      <td>4.805166e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>high_salary_low_balance</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.951000e-01</td>\n",
       "      <td>3.962975e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           count          mean           std           min  \\\n",
       "CustomerId               10000.0  1.569094e+07  7.193619e+04  1.556570e+07   \n",
       "CreditScore              10000.0  9.891430e-03  1.469634e-03  5.321825e-03   \n",
       "Age                      10000.0  3.892180e+01  1.048781e+01  1.800000e+01   \n",
       "Tenure                   10000.0  8.661834e-03  4.997513e-03  0.000000e+00   \n",
       "Balance                  10000.0  7.748754e-03  6.321456e-03  0.000000e+00   \n",
       "NumOfProducts            10000.0  1.530200e+00  5.816544e-01  1.000000e+00   \n",
       "HasCrCard                10000.0  7.055000e-01  4.558405e-01  0.000000e+00   \n",
       "IsActiveMember           10000.0  5.151000e-01  4.997969e-01  0.000000e+00   \n",
       "EstimatedSalary          10000.0  8.670720e-03  4.982078e-03  1.003164e-06   \n",
       "Exited                   10000.0  2.038000e-01  4.028421e-01  0.000000e+00   \n",
       "Complain                 10000.0  2.044000e-01  4.032827e-01  0.000000e+00   \n",
       "Satisfaction Score       10000.0  3.013800e+00  1.405919e+00  1.000000e+00   \n",
       "Card Type                10000.0  1.498000e+00  1.118356e+00  0.000000e+00   \n",
       "Point Earned             10000.0  9.371039e-03  3.490681e-03  1.838625e-03   \n",
       "Female                   10000.0  4.543000e-01  4.979320e-01  0.000000e+00   \n",
       "Male                     10000.0  5.457000e-01  4.979320e-01  0.000000e+00   \n",
       "France                   10000.0  5.014000e-01  5.000230e-01  0.000000e+00   \n",
       "Germany                  10000.0  2.509000e-01  4.335527e-01  0.000000e+00   \n",
       "Spain                    10000.0  2.477000e-01  4.316982e-01  0.000000e+00   \n",
       "female_male_ratio        10000.0  8.325087e-01  2.116191e-13  8.325087e-01   \n",
       "tenure_balance_ratio     10000.0  4.892430e-01  5.755533e-01  0.000000e+00   \n",
       "salary_balance_ratio     10000.0  5.007213e-01  7.014547e-01  0.000000e+00   \n",
       "high_balance_low_salary  10000.0  3.617000e-01  4.805166e-01  0.000000e+00   \n",
       "high_salary_low_balance  10000.0  1.951000e-01  3.962975e-01  0.000000e+00   \n",
       "\n",
       "                                  25%           50%           75%  \\\n",
       "CustomerId               1.562853e+07  1.569074e+07  1.575323e+07   \n",
       "CreditScore              8.879845e-03  9.913800e-03  1.091734e-02   \n",
       "Age                      3.200000e+01  3.700000e+01  4.400000e+01   \n",
       "Tenure                   5.183830e-03  8.639716e-03  1.209560e-02   \n",
       "Balance                  0.000000e+00  9.847145e-03  1.293159e-02   \n",
       "NumOfProducts            1.000000e+00  1.000000e+00  2.000000e+00   \n",
       "HasCrCard                0.000000e+00  1.000000e+00  1.000000e+00   \n",
       "IsActiveMember           0.000000e+00  1.000000e+00  1.000000e+00   \n",
       "EstimatedSalary          4.418263e-03  8.679701e-03  1.294136e-02   \n",
       "Exited                   0.000000e+00  0.000000e+00  0.000000e+00   \n",
       "Complain                 0.000000e+00  0.000000e+00  0.000000e+00   \n",
       "Satisfaction Score       2.000000e+00  3.000000e+00  4.000000e+00   \n",
       "Card Type                0.000000e+00  1.000000e+00  2.000000e+00   \n",
       "Point Earned             6.334757e-03  9.347630e-03  1.237595e-02   \n",
       "Female                   0.000000e+00  0.000000e+00  1.000000e+00   \n",
       "Male                     0.000000e+00  1.000000e+00  1.000000e+00   \n",
       "France                   0.000000e+00  1.000000e+00  1.000000e+00   \n",
       "Germany                  0.000000e+00  0.000000e+00  1.000000e+00   \n",
       "Spain                    0.000000e+00  0.000000e+00  0.000000e+00   \n",
       "female_male_ratio        8.325087e-01  8.325087e-01  8.325087e-01   \n",
       "tenure_balance_ratio     0.000000e+00  3.027190e-01  8.620636e-01   \n",
       "salary_balance_ratio     0.000000e+00  3.186064e-01  8.764975e-01   \n",
       "high_balance_low_salary  0.000000e+00  0.000000e+00  1.000000e+00   \n",
       "high_salary_low_balance  0.000000e+00  0.000000e+00  0.000000e+00   \n",
       "\n",
       "                                  max  \n",
       "CustomerId               1.581569e+07  \n",
       "CreditScore              1.292443e-02  \n",
       "Age                      9.200000e+01  \n",
       "Tenure                   1.727943e-02  \n",
       "Balance                  2.541838e-02  \n",
       "NumOfProducts            4.000000e+00  \n",
       "HasCrCard                1.000000e+00  \n",
       "IsActiveMember           1.000000e+00  \n",
       "EstimatedSalary          1.732515e-02  \n",
       "Exited                   1.000000e+00  \n",
       "Complain                 1.000000e+00  \n",
       "Satisfaction Score       5.000000e+00  \n",
       "Card Type                3.000000e+00  \n",
       "Point Earned             1.545063e-02  \n",
       "Female                   1.000000e+00  \n",
       "Male                     1.000000e+00  \n",
       "France                   1.000000e+00  \n",
       "Germany                  1.000000e+00  \n",
       "Spain                    1.000000e+00  \n",
       "female_male_ratio        8.325087e-01  \n",
       "tenure_balance_ratio     9.566742e+00  \n",
       "salary_balance_ratio     4.017499e+01  \n",
       "high_balance_low_salary  1.000000e+00  \n",
       "high_salary_low_balance  1.000000e+00  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eng_data.describe().T"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19ae729a",
   "metadata": {},
   "source": [
    "# Splitting Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "67bad2c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV, StratifiedKFold\n",
    "\n",
    "train, test = train_test_split(eng_data, test_size=0.2,random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a0375918",
   "metadata": {},
   "outputs": [],
   "source": [
    "features = eng_data.drop(['CustomerId','Surname','Exited'], axis=1).columns.to_list()\n",
    "target = 'Exited'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "49867735",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = train[features].astype('float')\n",
    "y_train = train[target].astype('float')\n",
    "X_test = test[features].astype('float')\n",
    "y_test = test[target].astype('float')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "e5eac5ff",
   "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>CreditScore</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>HasCrCard</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>Complain</th>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <th>...</th>\n",
       "      <th>Male</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Spain</th>\n",
       "      <th>zero_balance</th>\n",
       "      <th>female_male_ratio</th>\n",
       "      <th>tenure_balance_ratio</th>\n",
       "      <th>salary_balance_ratio</th>\n",
       "      <th>high_balance_low_salary</th>\n",
       "      <th>high_salary_low_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9254</th>\n",
       "      <td>0.010431</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.010368</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.015515</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1561</th>\n",
       "      <td>0.009610</td>\n",
       "      <td>42.0</td>\n",
       "      <td>0.006912</td>\n",
       "      <td>0.012119</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.016977</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.570320</td>\n",
       "      <td>1.400882</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1670</th>\n",
       "      <td>0.008500</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0.005184</td>\n",
       "      <td>0.011624</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.007441</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.445949</td>\n",
       "      <td>0.640097</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6087</th>\n",
       "      <td>0.008530</td>\n",
       "      <td>27.0</td>\n",
       "      <td>0.015551</td>\n",
       "      <td>0.013741</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.013261</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>1.131731</td>\n",
       "      <td>0.965060</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6669</th>\n",
       "      <td>0.007861</td>\n",
       "      <td>56.0</td>\n",
       "      <td>0.015551</td>\n",
       "      <td>0.014401</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.003421</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>1.079898</td>\n",
       "      <td>0.237542</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5734</th>\n",
       "      <td>0.011678</td>\n",
       "      <td>54.0</td>\n",
       "      <td>0.013824</td>\n",
       "      <td>0.007063</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.006010</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>1.957297</td>\n",
       "      <td>0.851024</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5191</th>\n",
       "      <td>0.010370</td>\n",
       "      <td>58.0</td>\n",
       "      <td>0.001728</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000061</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5390</th>\n",
       "      <td>0.011176</td>\n",
       "      <td>38.0</td>\n",
       "      <td>0.001728</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.007989</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>860</th>\n",
       "      <td>0.010142</td>\n",
       "      <td>43.0</td>\n",
       "      <td>0.013824</td>\n",
       "      <td>0.019272</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.008447</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.717291</td>\n",
       "      <td>0.438309</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7270</th>\n",
       "      <td>0.010598</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.001728</td>\n",
       "      <td>0.014985</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.004642</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.115314</td>\n",
       "      <td>0.309761</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8000 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      CreditScore   Age    Tenure   Balance  NumOfProducts  HasCrCard  \\\n",
       "9254     0.010431  32.0  0.010368  0.000000            2.0        1.0   \n",
       "1561     0.009610  42.0  0.006912  0.012119            2.0        1.0   \n",
       "1670     0.008500  24.0  0.005184  0.011624            1.0        1.0   \n",
       "6087     0.008530  27.0  0.015551  0.013741            1.0        1.0   \n",
       "6669     0.007861  56.0  0.015551  0.014401            1.0        0.0   \n",
       "...           ...   ...       ...       ...            ...        ...   \n",
       "5734     0.011678  54.0  0.013824  0.007063            1.0        1.0   \n",
       "5191     0.010370  58.0  0.001728  0.000000            1.0        1.0   \n",
       "5390     0.011176  38.0  0.001728  0.000000            3.0        0.0   \n",
       "860      0.010142  43.0  0.013824  0.019272            1.0        1.0   \n",
       "7270     0.010598  51.0  0.001728  0.014985            1.0        1.0   \n",
       "\n",
       "      IsActiveMember  EstimatedSalary  Complain  Satisfaction Score  ...  \\\n",
       "9254             1.0         0.015515       0.0                 2.0  ...   \n",
       "1561             1.0         0.016977       0.0                 4.0  ...   \n",
       "1670             0.0         0.007441       1.0                 4.0  ...   \n",
       "6087             0.0         0.013261       1.0                 2.0  ...   \n",
       "6669             0.0         0.003421       1.0                 3.0  ...   \n",
       "...              ...              ...       ...                 ...  ...   \n",
       "5734             1.0         0.006010       0.0                 3.0  ...   \n",
       "5191             1.0         0.000061       0.0                 4.0  ...   \n",
       "5390             0.0         0.007989       1.0                 5.0  ...   \n",
       "860              0.0         0.008447       1.0                 3.0  ...   \n",
       "7270             1.0         0.004642       0.0                 4.0  ...   \n",
       "\n",
       "      Male  France  Germany  Spain  zero_balance  female_male_ratio  \\\n",
       "9254   1.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "1561   1.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "1670   1.0     0.0      0.0    1.0           0.0           0.832509   \n",
       "6087   0.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "6669   1.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "...    ...     ...      ...    ...           ...                ...   \n",
       "5734   1.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "5191   0.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "5390   0.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "860    1.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "7270   1.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "\n",
       "      tenure_balance_ratio  salary_balance_ratio  high_balance_low_salary  \\\n",
       "9254              0.000000              0.000000                      1.0   \n",
       "1561              0.570320              1.400882                      0.0   \n",
       "1670              0.445949              0.640097                      0.0   \n",
       "6087              1.131731              0.965060                      0.0   \n",
       "6669              1.079898              0.237542                      0.0   \n",
       "...                    ...                   ...                      ...   \n",
       "5734              1.957297              0.851024                      0.0   \n",
       "5191              0.000000              0.000000                      1.0   \n",
       "5390              0.000000              0.000000                      1.0   \n",
       "860               0.717291              0.438309                      0.0   \n",
       "7270              0.115314              0.309761                      0.0   \n",
       "\n",
       "      high_salary_low_balance  \n",
       "9254                      0.0  \n",
       "1561                      1.0  \n",
       "1670                      0.0  \n",
       "6087                      0.0  \n",
       "6669                      0.0  \n",
       "...                       ...  \n",
       "5734                      0.0  \n",
       "5191                      0.0  \n",
       "5390                      0.0  \n",
       "860                       0.0  \n",
       "7270                      0.0  \n",
       "\n",
       "[8000 rows x 23 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "id": "4959b256",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training samples: 8000\n",
      "Number of test samples: 2000\n"
     ]
    }
   ],
   "source": [
    "print('Number of training samples:', X_train.shape[0])\n",
    "print('Number of test samples:', X_test.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79e34581",
   "metadata": {},
   "source": [
    "# Building Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "ffa3c8c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "5180eba6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.base import BaseEstimator, TransformerMixin\n",
    "\n",
    "class custom_scaler(BaseEstimator, TransformerMixin):\n",
    "    def __init__(self, variables=None, excludeComplain=None):\n",
    "        self.variables = variables\n",
    "        self.excludeComplain = excludeComplain\n",
    "        \n",
    "    def fit(self, X, y=None):\n",
    "        return self\n",
    "    \n",
    "    def transform(self, X):\n",
    "        scaler = StandardScaler()\n",
    "        num_cols = [col for col in X.columns if col not in self.variables]\n",
    "        if (self.excludeComplain) & ('Complain' in self.variables):\n",
    "            self.variables.remove('Complain')\n",
    "        X_scaled = pd.DataFrame(scaler.fit_transform(X[num_cols]), index=X.index,columns=num_cols)\n",
    "        X = pd.concat([X_scaled, X[self.variables]], axis=1)\n",
    "        return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "26fec19c",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_cm(y_hat):\n",
    "    matrix = confusion_matrix(y_test,y_hat)\n",
    "    ax = sns.heatmap(matrix, annot=True, fmt='d', cmap='crest')\n",
    "    ax.set(title='Confusion Matrix', xlabel='Predicted values', ylabel='Actual Values',\n",
    "           xticklabels=['Did not churn', 'Churn'], yticklabels=['Did not churn', 'Churn'])\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "3a2f1511",
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_model(model_name, parameters, excludeComplain=False):\n",
    "    cat_cols = ['HasCrCard', 'IsActiveMember', 'Complain', 'Card Type', 'Female', 'Male', \n",
    "            'France','Germany','Spain','zero_balance','high_balance_low_salary','high_salary_low_balance']\n",
    "    scale_num_cols = custom_scaler(variables=cat_cols, excludeComplain=excludeComplain)\n",
    "    pipe = Pipeline([('scaler', scale_num_cols),('model',model_name)])\n",
    "    pipe_cv = GridSearchCV(pipe, param_grid=parameters, refit=True, verbose=1, scoring='roc_auc')\n",
    "    pipe_cv.fit(X_train,y_train)\n",
    "    return pipe_cv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "240c5872",
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model_name, excludeComplain=False):\n",
    "    features = X_test.columns.to_list()\n",
    "    if excludeComplain:\n",
    "        features.remove('Complain')\n",
    "    \n",
    "    print('The best parameters are:', model_name.best_params_) \n",
    "    yhat_train = model_name.predict(X_train[features])\n",
    "    yhat = model_name.predict(X_test[features])\n",
    "    train_acc = accuracy_score(yhat_train, y_train)\n",
    "    test_acc = accuracy_score(yhat, y_test)\n",
    "    pos_probs = model_name.predict_proba(X_test[features])[:,1]\n",
    "    roc_score = roc_auc_score(y_test, pos_probs)\n",
    "    print('The best score on training set is:', train_acc)\n",
    "    print('Accuracy on test set: ', test_acc)\n",
    "    print('ROC score: ', roc_score)\n",
    "    print('The classification scores are:')\n",
    "    print(classification_report(yhat, y_test))\n",
    "    scores['train_acc'].append(train_acc)\n",
    "    scores['test_acc'].append(test_acc)\n",
    "    scores['roc_auc'].append(roc_score)\n",
    "    plot_cm(yhat)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7def9e42",
   "metadata": {},
   "source": [
    "# Experiment with complain vs without complain as a feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "896defc6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialise score metrics\n",
    "scores = {'test_acc':[], 'train_acc':[], 'roc_auc':[]}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "df66f19d",
   "metadata": {},
   "source": [
    "## Logistic Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e4da5467",
   "metadata": {},
   "source": [
    "### With Complain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "id": "e59f4cc5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 6 candidates, totalling 30 fits\n"
     ]
    }
   ],
   "source": [
    "parameters = {'model__C':[0.01,0.1,1], \n",
    "              'model__penalty':['l2'], \n",
    "              'model__solver':['lbfgs','newton-cholesky']}\n",
    "LR_model = build_model(LogisticRegression(random_state=42), parameters, excludeComplain=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 373,
   "id": "6ec1749c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are: {'model__C': 0.01, 'model__penalty': 'l2', 'model__solver': 'lbfgs'}\n",
      "The best score on training set is: 0.99775\n",
      "Accuracy on test set:  0.999\n",
      "ROC score:  0.9993017191010702\n",
      "The classification scores are:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       1.00      1.00      1.00      1607\n",
      "         1.0       1.00      1.00      1.00       393\n",
      "\n",
      "    accuracy                           1.00      2000\n",
      "   macro avg       1.00      1.00      1.00      2000\n",
      "weighted avg       1.00      1.00      1.00      2000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate(LR_model)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e852e75d",
   "metadata": {},
   "source": [
    "### Without Complain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 375,
   "id": "6b00d03a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 6 candidates, totalling 30 fits\n"
     ]
    }
   ],
   "source": [
    "parameters = {'model__C':[0.01,0.1,1], \n",
    "              'model__penalty':['l2'], \n",
    "              'model__solver':['lbfgs','newton-cholesky']}\n",
    "LR_model2 = build_model(LogisticRegression(random_state=42), parameters, excludeComplain=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 376,
   "id": "e3c7eed8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are: {'model__C': 0.01, 'model__penalty': 'l2', 'model__solver': 'lbfgs'}\n",
      "The best score on training set is: 0.806\n",
      "Accuracy on test set:  0.812\n",
      "ROC score:  0.7810928966940122\n",
      "The classification scores are:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.97      0.82      0.89      1895\n",
      "         1.0       0.16      0.58      0.24       105\n",
      "\n",
      "    accuracy                           0.81      2000\n",
      "   macro avg       0.56      0.70      0.57      2000\n",
      "weighted avg       0.93      0.81      0.86      2000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate(LR_model2, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "01286443",
   "metadata": {},
   "source": [
    "We can see that the model performance dropped significantly when Complain is not included as a one of the features. We know that Complain is one of the key drivers for churning customers. Thus, now we want to find out whether there is other customer segments that drives churning customers. Now we are going to exclude Complain as a feature.  \n",
    "\n",
    "Logistic Regression is our base model. It is a good indicator whether other classifiers can improve from Logistic Regression and make comparison."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a65c031a",
   "metadata": {},
   "source": [
    "## Decision Tree Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 377,
   "id": "3b430a94",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 24 candidates, totalling 120 fits\n"
     ]
    }
   ],
   "source": [
    "parameters = {'model__min_samples_split':[2,8],\n",
    "              'model__max_depth':[2,4,8], \n",
    "              'model__max_features':[5,10],\n",
    "              'model__class_weight':[{0:1,1:2}, {0:1,1:1.5}]}\n",
    "tree_model = build_model(DecisionTreeClassifier(random_state=42), parameters, excludeComplain=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 378,
   "id": "c5e94917",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are: {'model__class_weight': {0: 1, 1: 2}, 'model__max_depth': 2, 'model__max_features': 5, 'model__min_samples_split': 2}\n",
      "The best score on training set is: 0.745625\n",
      "Accuracy on test set:  0.7565\n",
      "ROC score:  0.732778508782347\n",
      "The classification scores are:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.79      0.90      0.84      1416\n",
      "         1.0       0.62      0.42      0.50       584\n",
      "\n",
      "    accuracy                           0.76      2000\n",
      "   macro avg       0.71      0.66      0.67      2000\n",
      "weighted avg       0.74      0.76      0.74      2000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate(tree_model, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "062e2c6f",
   "metadata": {},
   "source": [
    "## Random Forest Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f4e27f37",
   "metadata": {},
   "source": [
    "Ensemble models can boost performance by combining multiple algorithms to learn the data. Random Forest is one of example of Ensemble models. It uses bagging technique which uses multiple decision trees."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 383,
   "id": "ff94978f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 24 candidates, totalling 120 fits\n"
     ]
    }
   ],
   "source": [
    "parameters={'model__n_estimators':[100,200], \n",
    "            'model__class_weight':[{0:1,1:2}, {0:1,1:1.5}],\n",
    "            'model__max_depth':[2,4,8],\n",
    "            'model__max_features':[5,15]}\n",
    "rf_model = build_model(RandomForestClassifier(random_state=42), parameters, excludeComplain=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 384,
   "id": "ae3ea413",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are: {'model__class_weight': {0: 1, 1: 2}, 'model__max_depth': 2, 'model__max_features': 5, 'model__n_estimators': 100}\n",
      "The best score on training set is: 0.8355\n",
      "Accuracy on test set:  0.8405\n",
      "ROC score:  0.8249476289325803\n",
      "The classification scores are:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.99      0.84      0.91      1900\n",
      "         1.0       0.22      0.87      0.35       100\n",
      "\n",
      "    accuracy                           0.84      2000\n",
      "   macro avg       0.61      0.85      0.63      2000\n",
      "weighted avg       0.95      0.84      0.88      2000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate(rf_model, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41990e88",
   "metadata": {},
   "source": [
    "## Gradient Boosting"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5344d2e9",
   "metadata": {},
   "source": [
    "Gradient boosting is another ensemble model that utilises boosting technique. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 388,
   "id": "b5bb033e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 16 candidates, totalling 80 fits\n"
     ]
    }
   ],
   "source": [
    "parameters = {'model__n_estimators':[100,150], \n",
    "              'model__learning_rate':[0.1,0.2],\n",
    "              'model__max_features':[5,10],\n",
    "              'model__subsample':[1,0.9]}\n",
    "gb_model = build_model(GradientBoostingClassifier(random_state=42), parameters, excludeComplain=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 389,
   "id": "3a1cf274",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are: {'model__learning_rate': 0.1, 'model__max_features': 5, 'model__n_estimators': 100, 'model__subsample': 1}\n",
      "The best score on training set is: 0.87025\n",
      "Accuracy on test set:  0.8605\n",
      "ROC score:  0.8683819675687316\n",
      "The classification scores are:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.96      0.88      0.92      1760\n",
      "         1.0       0.45      0.74      0.56       240\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.71      0.81      0.74      2000\n",
      "weighted avg       0.90      0.86      0.87      2000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate(gb_model, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c6265b9",
   "metadata": {},
   "source": [
    "## Model Performance Comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 393,
   "id": "ac182bfa",
   "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>test_acc</th>\n",
       "      <th>train_acc</th>\n",
       "      <th>roc_auc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Logistic Regression</th>\n",
       "      <td>0.8120</td>\n",
       "      <td>0.806000</td>\n",
       "      <td>0.781093</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Decision Tree1</th>\n",
       "      <td>0.7565</td>\n",
       "      <td>0.745625</td>\n",
       "      <td>0.732779</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest</th>\n",
       "      <td>0.8405</td>\n",
       "      <td>0.835500</td>\n",
       "      <td>0.824948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gradient Boosting</th>\n",
       "      <td>0.8605</td>\n",
       "      <td>0.870250</td>\n",
       "      <td>0.868382</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     test_acc  train_acc   roc_auc\n",
       "Logistic Regression    0.8120   0.806000  0.781093\n",
       "Decision Tree1         0.7565   0.745625  0.732779\n",
       "Random Forest          0.8405   0.835500  0.824948\n",
       "Gradient Boosting      0.8605   0.870250  0.868382"
      ]
     },
     "execution_count": 393,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scores_df = pd.DataFrame(scores, index=['Logistic Regression', 'Decision Tree', 'Random Forest', 'Gradient Boosting'])\n",
    "scores_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 394,
   "id": "6112358a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0.7, 0.9)]"
      ]
     },
     "execution_count": 394,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = scores_df.drop('roc_auc',axis=1).plot(kind='line', rot=0, figsize=(10,6))\n",
    "ax.set(ylim=[0.7,0.9])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 395,
   "id": "7590413d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = scores_df['roc_auc'].plot(kind='line', rot=0, figsize=(10,6))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e86b150",
   "metadata": {},
   "source": [
    "There is a small improvement in validation score from logistic regression model. Ensemble methods are great here. Random Forest and Gradient Boosting models have significant increase in performance. The ROC score also increases from logistic regression and decision tree.  \n",
    "\n",
    "In conclusion, one of the main driver of churning customer is complain. We know this finding based on our exploration and further proved it with building ML models. When building models without complain as a feature, performance dropped significantly. So, I decided to investigate whether there are other features that interact with churn well. \n",
    "\n",
    "Several models were built including Logistic Regression (our base model), decision tree, random forest, and gradient boosting. Overall, gradient boosting model is chosen here as the best model for predicting churn. While random forest has slightly higher test accuracy, gradient boosting has better ROC score. ROC score is a measure to evaluate whether model can distinguish binary classes. \n",
    "\n",
    "The main features that drive customer churn will be discovered in the next phase using recursive feature elimination (RFE) and cross validation (CV). I will also experiment using resample method to find out if it will improve the f1-score for `1` class (churn)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "72332584",
   "metadata": {},
   "source": [
    "Current model performance benchmark (GB):"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjoAAAElCAYAAADtIEgTAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAEcWSURBVHhe7Z07luq6Eoar7yggYxiQMQM6NNkO9xBgBDCEE+4MwmYGZDAMMpgFV6WHractG9NNm/9bS+fsNkKv0qNUKuGPu4CGynlLH7O1+vfmRPfVVP0bAAAAAG/BsBUdAAAAALw1/9P/BwAAAAAYHFB0AAAAADBYoOgAAAAAYLBA0QEAAADAYIGiAwAAAIDBAkUHAAAAAIOltaJz23/Sx8cn7W/6wTdz3n6I/Ld01n8DMAzOtP3gvv1B29469432nyrNz58asEODf5tLyglzEAC/BaHoVJNhOvycYvOS6MmuvwUJ/Cx6DHzuxb/AIHjrMRrO6Y+3Q6WIl8FP9LanTz+OE8J1RG2crTiJggbxEopmGK8+PngPhKIzouLrTvy7gTJcd7QQHyx21+rZ/YuKkfoCAOAZTGmlx1t/P+Bdje0vDOB+mK70nLgSEntFWCEZ05J2dNX96X7a0Hr2iFWP05zR2VkTTrRZz9zNwaigr/JzO1xpJxeVguZWN2Tr/HhJtLuaeJE0Bay8jPdFVR+OR2uaJZWXDZ3KuCa8qrzAdwAfHQAAGAi3/VaoAGKh/yqEmqsRypnQdeiw/NfJqmHSXDnK8pT+sPZy2NOxSX86/6PlQagfK6tMtz1t1/xmHnsTLZR9VVD6ZxV0VHzR3a6PiSdKtcVRA8jgIUVH+cvokDT7hybPfszJnnk2lWh5pm5CahcQmnvdHZBVD/3+LN4lpeO3pU07NZXVEMaLtZOUo5SfHz/WVrl5tyPHNF2V02urjh2qylPsgMVEzBPs2E7XK0O7dtLxa9KTBOb+xDGxjsdVddsqFj+3L6l4Un7eOKmLb8crQ0cZMIHsO88l1uc9j9H8MiraxTfljvejfG503LNGMfesFze6yITXdOyQweXCg+NMl05NJ8YKazRCUZpbhbod93TwnjE3VVBaNxV0MpEnDwBkcfe57u6iA90Xu6t+4HLdLe5Ei/tiYcc53YV+fRfquf5bo9Nynp82/G6tZPpNCEX+TguRv0ijTDaWj0CV1YonkN8X5Xez1+Vf7O7249PGj6fRdfCr25lo+a/33WIjSuZh4jaVNZZm8rviGT8XIZCpEzfyLNH2+XA9OW+7ruaZWye7nGV2Ov+u/Umh8/PaxSe/nVRct0l0PKeeLmZsRati2tnLS5UpnaaIIL8TF48pk5tGtBwROZvx9Ujbq/J7eYkyB2nG+lndXNL3GC1p7ivZdTLosj5c3uhYUDLenNT/O8nKlM8pXLzfB0TLFG9D2W4iD/n/pnQTsld9smY8gLeko6ITfh6bcFMdNqsjJ1D5hBNC2MFTAzvyvKHOAX1Poi3Si7VzjGQbR/JKtamfV2oSeWhykeWJLe56IrUKFS9nnpJSTztFp6mdkiTrqlDtmPjcLPR+GZv6Tu3nuo39skfGQ7xsoYzakS872caReKnnje3SmaYyd+mPCTm0JZCbSle1wWOyMn1ffV/XsaYvG+JjI2wjGU+XLf4dm3T+qp/yZ1ZoJQswRDoeXS2osL3KopzpuBYxi7l1tqqYTMRwzDnbTRKaPEcTfmCZV89HWkfLOSGV/bEyJY8m0tR7WI57OYZpjTbDspm9/gRAt+nuj2ea9km3PU3novVEjMAOHbbpdGU78aXM4qbtu5nFz6qgjpOiYkpzVdBKTpKwnN9LUzs1caDLRf+zA1GZPkpEpt/DiGTX4SPDhqOd580lfZNbJxvjiN6nwywfh82ITv04t8s+LjQQMUmROuplh9+GSyraD6dpvuIj3hmd6J5Z0PNWHzVvVkH+0p/HcUI+0UYeSePm8DvzPGfk20UMNaU8OGfVIoxlL30G1SKiznoPtBy7eZf+GA480aibAXZ5v03pkbcV+CaBUnZM/oHSo9u0kYx4h9ar7YXkUb2c6KoyyqD9IdpjfAdqOIh89T9/FxF/ls7t9POM5oVQxsV4srxElZMq60ndV1JeQMUunAde5R/lKwg/Mpd0J6tOT+WYVHIWYlPVHt2XZZJKgbjuzjTjZzU7s/O/pfTDcZ2YXY4pJWchNqT6nzbs+ySH0WJH1yzFSMzt8iax23fBe/E8RUdbSRbOlUQ79HtlXSk2CzaOSJSVYWFdXfSC48XP2NfsLaXnASfLdphdHYdK6emkbOm2j6IXjfYTnrKE0UZMSmU53dB+56h3v1G0EpSY8F4a6Tg8ozVPxnYb8Y74tyIUY6lSWIquuRrcXu4u9i68UhAsx9xvnkv6oLFOz6C0TAvB+EqOtHCLWWbSvqHOW9GXxYx0sixOXD9ZL9EfolOktubErYWV1WsZKDnKeicK6s3PnOSnUmx5XAXzNwBpnni9PHJE9ETkzQD7CEQeBx1o38mmzUqP/u2H4OjkO2ClRyk7juVlNKciq01r2l4uWDlHjz56clofe52s00cPyoL0lKOaZ6OVAuc67S9HHTF6itsTFIxyARW9rDpd/d65pG/idbLp69aVPu4VM4dvZJPyizxnyhtiUatThtU1grLmiI3mn0iGgqkqaGgN1ApZ+HyrrXfe1fkczHh8wPIIfjdPVHSEsrASnZl3Mk+2ihhzprOwjApS2Y8b/F4EfLXWj3Q7Eruk1Pm5rLf9mKO5/IHlJjrg02163tpn0Il4bGlQDdVpgZr+YRPwmmY9muFHxUq0JR8x2pM8X99Wu8g6s3d/jGguNUj39zsexb4iKxeTX3x0Jen9GJEXeH9x175gjt9Wh7mk5zGaT26dLPRYF6Xt5OdmM13pH9Oz20nMb3LYn2I+QLpsTHTDYayuXppiLvmrFY9AfyitOTXzjP5tn/XMbivRdqqgnjVKzM9y7LhWpSzMd/00wXshdmYuWbeuQm/3tKe8uVHgha7XIfRtCjfUeOhH44e3MVS96uO4hPVKtVkOqv3sUHejIcw7WlZzU6chXlp2McyNBy88dLMhkmakoPFyhjc4uhL2ATevNu0UpMX10fKwqxbrd1Ww8kqNy9jtokSfV8HuV4mbONG8EuOYQ0RW2UT6aDq9tnNJX2M00edNSLRfbRwHU87cMdiEX+/621FlH6wZQ9F+moivxkl9ngZ/3gvl09D2Tj6xuHnlAMPmg/8jOgQAACRgC5u5aePtqK0dc+6tGQAA+E6eeHQFABgE+hg36liqfxoBAABeFSg6AIB6zC2+iCO6cjplHQjWHADAa4KjKwBABub4Sv9Z0sFBFAAAvhEoOgAAAAAYLDi6AgAAAMBggaIDAAAAgMECRQcAAAAAgwWKDgAAAAAGCxQdAAAAAAwWKDoAAAAAGCyBolO+yfbDfkkkGDz8U/5C7rnvTOwb1e9+QZ/74XbKh3/3hsdxFR4vs3nLthVqE80rQzXn6BCPRJ92HCu8viwAAD+Jp+ioN9kudjviN0rvw1fZAgBeHlZIxrSkHV3vd35xL/HbE9ezj/At+dlwmjM6764qPRlOtFnP6CP2NnupEIoyTE9W/HvwBunz9oPGS6LdNSNNweZUpZVKEwAAbFxFR77TZkHFfE6TBb+1/xidbAAAr8ttv6U1/2LxV0Ej/YymKxK6Dh2W/4LXOORg0lwVZYqCKf3ZyYmCnD0RW19mIjYrJXVaiIi3le8D/aIq2SmtVEHpHyw1AIAecBSd23FPh0VB89GI5kVkArMJTMk1xw7a3F+FrTvZ6rSC3WbsmEA+U3n5Ju8ss3hdORvqxLvP1PfrPssjPBYI2kOXj+vp1qtrvlae/AZqAe/6q3RjFoDwOCJtJQjr5MjIau+xfLfAgZbjmvhtaep3JU116tJOCiOndBv1jbLKhi/gvNFFVn5Nxw5termwfM50yaiGmkd29KfB0iLjCeXJf03WTRWU1l0KCgAAPveS611szu68DVN/7u5C1bmbP22uuwW/NuK+2F31E+Z03yx2IhUXsTmTcZ10OG37gc7LTU9w2oTf1c9ksPJT+SzudhJcTjdNXUcvHpNVJ513UE6Ox+WJNVYOsXRNPSPtFK/7RpTiAWJtHaDrabeJKZP/xehzbv94OVX7h3LpipGnm73X7yQt6sRktZPB9Dcv/Weiyx70Y1lm9f+w/2YQ64+xtjN1bmwgHc9rF9mXxXfl/yMyyWt3AACoqBSdYCJJTVh6csuZcaKTboRUvNiiEp1wBbkLUHTCzK1TfHJWeXddpNMLQ7D467LH839wEchIQ5UnVFSiz1uWqW9FJ1f5a1UnplO9OioXXQjGklFyqn9njd0Iqk3N93W/DWSm8uD8y/gmOPmGY8koOeW/7fY3fd8J/fUXAMBwKY+uQjOyPr5aH11z//lIa1rQrskuLTBpuuf6/bDx7d3TVSvHxLNtg8+uU+xI70Z7djSQR376URukX1SkPoLRvBClOtDloh9oFsW88r34NlJHIqKcE37iHYlMJqLs6ojnoeOnjkzYyUyUaZY8rmJa1qkDo+JL9suvJ4yBZpQDMZ36cdidrsTGSGggQqgkHY0P/OZy27+m4rAc03FeOQwLRYUWNU7GfPQ7o1Pap2dU0JdJS4fTRh11ft+xIADgN6IVnfiEH5vszfl5Dupc/ycJfS8+xkuhOri0qdOoWAnVTUywxlPSKCory/GzDUKLaWolRyn7MS4kxSkXOa9Ntd+Kg1yYTqKt+CtV3O9SeqSCwYuyVHZM/r7S07JOv4pjUslZCCW0PdpPSSapFI3r7qzaNiLUxe7q5iv6w387f5OgOKaUnMWE6ko6XV1JJtnRwRoA8B4oRUcv1sGEbxwvf6VTIE/MZtdp7QR5Z6ljdGNKc7l+KktXyqEyG235iKKVoOnkJ6wBPmLR4YJu3OvCdgg341NalZ9XSs+37cC1lU+GUumxHbe71OnFGU3kZuWwXIZKjrRcdutP5+1MfJfH0qrcDLEyeWVNQ8wbla6j2vTgmyEFauNkLJQjUn8uaRkoOWc6qoL+gOUSADA0pKJjFmtHIdBBrQ/V8ZU5Tsn5jZ2p0gg6m//PcrbryO0iy7zY/XGsVDHa1ImZ/mFlietlfneoOY8kemGKKZPK0vSAEtUremHyjzKzYaVHKTuxRfDpsNIjO7N9FPhonZr5/ltXWhGP9Bs1nuL9qbzFFz1aMje2ctBter4E6fj9Wc0P4klYUKmQxY5zXbRFLnL0CAAAJTGnQIeI86VYL4Jn4mnk1lXCYZEdC3McE7XToRM11xk05rirv8vBdw7NrxNj6sWh2eG1iajDqi6r8yxwNNW0dJCNox1VU/2AibVpAq5Tq3KatB+rhIblEzqqKhl78mpRJ0VGO5VY/SQ7/T7QZbTbsraP2P057uBbjkc7AdN2iTbN6buhTCJlj6Lj9TD+AADDho9yohNQRWLisZQGFdITjlnIyxCd9M3EpQPnp8vmZN1mUffLKPNV+UTr26JOZdysgmQQ5B1ZcFKyatMmtXgyiOXlLIpWiMjUVlZVaLglo+tnf6d7ncK6pJWN/DopctpJ8e23rkr8Mta3fTlGk3W24tghFb+FLP1+EmurWN7f36YAgN/IB/9HTBqgLfxjdDP166+/zo8DAAAAeBOg6HSCb3PpdwnZP7MPAAAAgJcCik4H5G9+rBe0u8Z/QwQAAAAArwEUnWzUj6+pe2DuNVsAAAAAvCZQdAAAAAAwWJy3lwMAAAAADAkoOgAAAAAYLFB0AAAAADBYoOgAAAAAYLBA0QEAAADAYIGiAwAAAIDBohUd/o2YD/X2YidsE2915l8GduPWvp2ZX5dgxeWwfdbrokE3fBlF32LdktuePu00RWjVT2JlyEyTf9TRjiNDstNF+n9NBy3f9J0cHwq/DPVvMA/HlF2EaH3s4JQ3NZ5ryhC0a6xuGe0UkY8bPqkqQlhnDrXtZKVfIyJFVp0AAIOHf0enfAGg89Y981JA/2WAkbjmBX6Rt/aZF/a5H/FLFPHW4VfBvDCxkpF5yeUDMoq9kDXZT0x+eS/9zEvTIxlP9Wf3BZG6j/svrNQvT13sTg3tE/t+ZNwYHnlBbPSFrrE6pYmPUZ8W7RRFy7gpbtnGfqwqr1PQX0Py6gQAeAdqFB1BZBKVE0hssorEDRdQ8HokFkWtGOQulj6pfqIWIFdBSPYpjzZpxojFU300/K56bilesj1MvHpFMJWmKICbJhNT3rJJKQ8JmUZIltUju51SRBWyGLE6qWfmuyrPdFq5dQIAvAe1Pjq3S2iW3q7FtFbMwxdZTuckJldaH813bnTcH0TkHf3p/V0JoQm9zowdmP2jkSNm9Eg8mZY8UvHjx83i1TGHDrVHQqZe32hiPx9pTQsq5p5ELxcS0qPD/lhT3hQ3kl3ncKGLelDDmY6iT21WTS9HbZNmDYsJTfQ/mcuFa3mmS1MlRwV9Zb72Q6bp5SOZTERLH7hpS27HPR26jpHzP1pyVrHxmIUao4vdn8Z6ZbdTFDFWeOIQM8Q8t57TiVWnERVfd1plfTe/TgCAN0HpOxGLjjH12zur2l2ZtxNLHhU8SDRdzju2g9P18nZ3vONzvx5JM1Z/gbIKqFDtmHU+0bj2ble3UWq3qds33cb9k7RyiLqoI4JuO2Oz63asCrF2jlk5EmSnGSPVd02bOx/E5VlRL8e0hcm3tOh0Ogo7P58Euu2ysu/UThqdT2N5BErG9f3B9INoudvUCQDwFriKjhf8ialpEnImXj0x5kxurUgtWBFyJk1Gljs2WUfyUnUM8w8WndSEWzsRGznEF9Bn4Jdbtplui9z2i2OUAdMH4nVTeYhnul04vgqxfPPSNKi0qzRTfdHIVC3iJo+6etcrOqbfuEqBX3ZGlZ//LstgQryDVNQqD/HxHCQpy8n1DOPHsm/fTopgbHj4dW+uupJrNF7LOgEAhk/comMWHW9maFr4nAlNT/apxaUz1oJYP3HpibhxdqsWm5Dws6ZJ21Au4PrvCq+tfxi7PrLMlsLXJO8czKIkQ0SZLD/3PjOLX6yZmtKMYxa+hOyMclIXp6RB0WGiipvfn0yZvHqa79bULbcfGso2szMq6+ymY+JGx0SrdhLUKmQRMupuyhcdQl3qBAAYNHEfnVFB/4mZnNZb6yqoeDzhU2/Xx6BC+1AY3wTpj8AuFQ95VIRIX4mT8geaVb4voTvNhaRbQRO3S6M/TJc6KJ+GNc2Mb04ZZuLp63HZf9J4X9D1y/eVmdKkiwOIvtor01QKNZ2mSxqLZ+H14Q2dvHynKy3j0udL0CpNnymtrjvRJ9e0deJqv6gZkVgYZZrX3VnJrfH+cg2yn6r0VPiiQvRJTnHqNahYfF3/EzP+Dns6xqplfOVa+KGMii8SylEwpkUqtLu6vkej4j9S2dv+Wd3a6fxvKWaMDa2KzE7E7cYFPSzpX+fmz60TAOAdSDojj4qVmJ4OtLRnG628nKMeiVqxME6Eo4maaM6XJ0wsYtEqF5BK6Wle7CKYcsbQStBC1Lstkwm3lFjAy3J6Ic+z8umocq5pGSg5xpk84lTbyI32f8UCt9g5aU5XQjGRa9jfcrFVynPMyVXkqzqb7j/5aSaJyPq8ZcWT5VQtjKwUXKWiP+v1956k47EYQVV3UnWMKdJ1mwqlPEQcyNsScY5WjEhlXzl+d2onrZDRZp4eYzH0PNOJFnUCALwHNbeupvRHTmLWDnA0p4In5tiuSN/e2ZXXR6Y0Zw3koZ1ZDqz0KGXHXTB0/uujVFbS6MUmVieRXtcFRS1Ua7INEs18/62r0byQi0pwc+d2JKnnRG/0VLfe4opApjWNSS5MKo0q/xZpppB9VPSM0qKirZDfwpn+8RWpzYoq44ZefCObAXXjMXJLqVQe7HRyMBbXgsrurBW/cOOi45YKSrd2MgpZNSfkESqELciuEwDgbbhLEn4j5rw8dq5vP0uewyd8EKSPQ75/gQ2ftQf56DIFZ/aJ835Ow4lbUye/TfJ9I/KdNUss/4egLk9E+S/Y5dRlT/lJWOUM+ozE1N3rE+Z7XrqxNg2ftUszoFae3nMTNylnU5bMPlxXxtjYiT3TqPK29ZtK98VQ9vFnrdsp0d6N6LaK1d2gypceI7l1AgC8B/WKjiA+sVYKjAl185mZmKrw2IRTTrpZ6VULpAnRSbSctKsQq5PKO3OBE4R155Aqr2nX/PT7Iihn7QKVUmBdQjmJkPhCEDehuOSlGfbPuj4SlVGQf9iPnODE9/Nv6O+Zfa+MVysbga2ImlDznbD+8f6X104KJaf29W6VdxnCfHLrBAAYPh/8HzERAAAAAAAMjtpfRgYAAAAA+M1A0QEAAADAYIGiAwAAAIDBAkUHAAAAAIMFig4AAAAABgsUHQAAAAAMFig6AAAAABgsUHQAAAAAMFig6AAAAABgsEDRAQAAAMBgaa3o3Paf9PHxWb3R/IVRZVVv2Zbhcx+8Jfq1aHoruEtQv6hcbrT/VJ9/vrjQyvq8vJwAAAD8FoSiUy2E6fA7FBufUfHFLy0V4Uq7hX44EFgpGC8P/K5GXUcOX1SMdISXQCtuOVobAAAA8ASEojOi4staLK87Yp1gsbu+8AI6VKa00m2+mupHUW503B9YSPSnNh5TyffrxYVYKqZfhSg1AAAA8Djw0fmVXOgi9ByaTqAQAAAAADU8pOict9bxVtKvovI7MaGXk4zz1knz42MrcnqEsJzJI5cWeWf5Cd329GnH6eWosH27O/LkEP1CcztV6cxozQ/WMze+1wZ5+RrC/EPfIxVHPvdk1dQGjbTqd+GxcNxPKqdOOi1ZAS9+dOzl5i3ofSwJgj7dR53CdqrvKrpeDwsdAPCruftcd/eFeLzYXfUDl+tucSda3BcLO87pvhHfYYcRB52W8/y0Ed9Pp5/DaSPSFGk42XFefv4l1/tOlJcWO/GvGFz+jfivhS6nXydV/7y8VTkXd6eqIt26upv2DaPoNq4LqfrruiSbp0zbbQMui/ud/HZSJPpFkob4sb4Tzd9uq6q86bbNo43s5XPO3+tzp03YH/LqpPuwrJNVhtgYM/W3847G48ct6pQLf9+rt8nH7fst6hR7Fms7CzNP+P0aAPBedFR0ws/VpOJOKPJZRLlIPc+ioXxxmhSdOLFyxuoZp1ueeYtxSwVCLwip6Hl5pknLs09FR7dn5LOw/EbR8eTUqe9U5Ms+N26bOhmlwJdT2M/Ud8O8Y8/b1OkxYnXNr1Oqj6X7niCqMAIA3o2OR1cLKuZN3iFnOq5FzGIe+JFMJmK5OezpmLCk13E77ulAG1p9l2Pt4UIX/U9Glp3WNGs0749owo7ChyWNk8d6r4B2bN6sHnM499qpd25HksWch97Xo3kheuSBLn4BNnNq9NVuQb7sdd/f/anPv0udFgW5Q087m5cO3EaeYd1HskOu6WgVPr9OPXG+hGOhsU4d55LpSjm313v2AwAGzvOckW8XOXEelmPnTJ0DX4vuykV64T6BwE/hg2bSwcRF3gwS20i1OJi48UViurqT2EUrZcfEfTmlRzs255LZTr0jVvymYp4vz23ZbNnrvt/IU+qk5en7RXGICKpNf25D+BtPY+o87J80lwAA3oPnKTqjidxRLpxr6nZ4nSvrclIWi4BfVjn/xzA7RQ7lIhF3IK5+y8dWer5p99wzrdupTyYT+bMHUbTCMJ18Q4dqIftGnlKnCUkjzebkyMgOgYGjzzoJ2Lk8/I2nB37L6hfNJQCA1+OJ18vVhHvYH3u1YEznaiK2ze+PoqxEHY/DeJGQi0PkmMGDlR6p7Ag158nGhxZMSTXpsVH5eqidHkUvduuI4G8XfrahyAnQc0nJfjSnIqfvP6VO+sg0Q55RWvTnODeSRc/6jadcOs4luHUFABA8UdEZUbESEyZbMPqcaKZ/5M5wPfN2nHyd9aF8bOVJXXsNLf38PNzpntmBIFiU+Cqsb7kxP/Tn+yT8LNM//CORYhfvHauxBSds0px2MhglatvZOlAxpT9K8O41ZbGYsfWg0R/mYdrIPt33z1s7jefUKSXPkDZ1aonjN8Nj4YGjq45ziaqHoKvSBwAYBnefrFtX4Q0dsQlM3N4wN2C88OBNCFUOK73g5oW50ZEI0dsfVeCPVR5+nSL1Sd36MFdi7RCpd1AXJ9S0aV0bmhsn0RDKL9YGYR9o004Vqm9YwWmvRP/QIShDUK9YXRLt8+Ctq2hZU7KXhPGjIsuqU3gTqZ5E/w++37ZOOfhpqvrIfuCk3bZOkbJySI0D3LoCAAg++D9iMgAAAAAAGBxPPLoCAAAAAPhZoOgAAAAAYLBA0QEAAADAYIGiAwAAAIDBAkUHAAAAAIMFig4AAAAABgsUHQAAAAAMFig6AAAAABgsUHQAAAAAMFig6AAgUe/t4pdAOu+dehD5xnd+sWTje6fAK1PK0YQaefLb2524fb7rrxVVny4D+iF4Q6DoAABAA6Pii98LKMJVvlS4jumK43E40UY/+35YyeEXqW7oJMuiw1dB/vuEKyXOfwkxAK8AvxRY9M8HNgxQdACQjKj4UovBV9Hfq+XLBTKywADwNG5H2h+Iat+Af95KBecvrRqVNwB+M1B0AABgaFwuJPQcmk4S6vVtT58zktaer2KiHwIwTKDogKcg/RSkP4A2O5oQmB+1H4F87sWN+hN4cUSos2g2+0u0S8/sgqsQN/e389MIyxD4CfHCJJ5zMq6/yCf16FL0zfy07JkwzXpZDYRRQV/3Vdra8xBt5BT6EcV95DLGiKCad/x0Y+M0N++fJ8dHTMUJ5wM1Duz6q7aUdfXmM19OKk3+rtdW0fHJ5Mgpb9xX43dGa36wnlXxvLiN3AF4AqcN3bl7cdiIbaPkursv+Fn5gLned4vMuLFnp4383mJ31Q8Mp/tGprkR/6q47hZVHj46rdTn/N3gcy5TMkFGlyMVJ1Z+/Sxadw6LnWg1hWpnt46/h5+WPcfz2i7W9g66zJYM0jTIvm9M2WtCvChGDj30o5icZPqRtE1cry1Pm8XdEWnuGBHY804VX8vBySfyLFr2n0fVKWwTv9+r+cmLJwjnCDM+3Oex+c08a25PQbacWox7yePjCIoOeApmwvH7pho49qAznd4foOGCItP0B5cg9jw16GvRgzI1nsIJI4e6QarrGPksKH9iUWgq82vzQrK3SOWlCMuW5vEJujOt+oWRQ9u+HaFFvnnjqcUYEag0w/z9vMJ5SJF6/nPk97dUvw/b2Sg6fj3DtlZpRjYTUs52Xm3kZPqbX9ZUXR8fRzi6Ak9kQ3PPNj6a8IMzXXyb46KgueNOoJ2DSyfeMx3XHG0eOPVOJkIFOOzpWKZ5oyN7Ym5W1KNfscqH1jTr63aKdhjd+I0kGM0LWtCBXS0cYvX/9byi7A8X8poe5DCZiH7LpwwNR8BGpnXO0kyHMRKbd9RNOHNUZ/rIPMhbzU9rOvYywPtgRLJIhyWN2xzV5BDUX+d1vnj5LKhwB6iWs9X2XeTUOO77A4oO+AFik1MDt4tULg7LsXtOK8J4KUaYg1ik/Ec9IG9Qie2RUnZM/g8oPaIRmop5DjTCN+SZsg98rj5oJh0CgE3gIyJDpO9L3x91rZ6VHRM3UHq0TBt5yhjRfcT3+eDwgsJnJe26Y4VeKDumnH0rPTYtlPyy7V98LoOiA76V24WntwVvCNoxmsjdx2J3Vde1g/DVq/UmyXRV5VkqPR0dgvXuN4qeOJK3Zt6JJ8leLt5iYfPTlWIFDtXvCNkh5cw8pVUZp1J6Ojn6PmWMTEgaZzcnqy5uWMUr9mPY7V8pPT1ZlktupKZn0T7qQRq/7V98LoOiA76VC2+lApNlDmpyOuyPGTuZKc3l7HrseSLwYKVHroodLFSMXsDXETu5UghDE/x78hzZy74o2nj1LRryu8JKj1J2DvYgGc2pyJHpU8aIPqJ59vzwJFjpkcqOKH2zkUQdEeahLF05x+NB27/4XAZFB3wbvINmy/Bm1eUMdkTFSkyXvJOpP/yXTP/sxA5jTTPPxMtlyPh6BL4SGbm6KWeRroN4Sn94wlrP3N3ueSuPZBr9F96GZ8re9sdQ115xdNUdbuPAcnM+ilYWo8QZJGmZnrf2OHvOGEn1kdeDr1/7lhvtY+RtGI0vzL50WOPv6qvZjXDf57gZin+07Z85l5nNy7b7T2ncAXgCytPfD7HbDPm3ChTmxoAXoh75xru/CvHbA26cKvi3AiJ5R8udKKMOzWUIb07c9a2r1HcfuJDwg/yw7CNxOCl1S8Tuq2E8Jzj5t5T9s2jsFw11ypZJSDj2I/25JGyvaJlzxohA5Z17ayrRBg/U/SmYG5d2SAnWa6d4f84fR+q7ftwaeWbJSbd7y3YO+lWL73/wf8SXAOgV/rGn2Zrfs/OsHyUDAADQHm3pYR+lBmckttCNl0S76zf5QD4JHF0BAAAAYLBA0QEAAADAYIGiAwAAAIDBAh8dAAAAAAwWWHQAAAAAMFig6AAAAABgsEDRAQAAAMBggaIDAAAAgMECRQcAAAAAgwWKDgAAAAAGCxQdAAAAAAwWKDoAAAAAGCxQdAAAAAAwWKDoAAAAAGCwQNEBAAAAwGCBogMAAACAwQJFBwAAAACDBYoOAAAAAAYLFB0AAAAADBYoOgAAAAAYLFB0AAAAADBYoOgAAAAAYLBA0QEAAADAYIGiAwAAAIDBAkUHAAAAAIMFig4AAAAABgsUHQAAAAAMFig6AAAAABgsUHQAAAAAMFig6AAAAABgsEDRAQAAAMBggaIDAAAAgMECRQcAAAAAgwWKDgAAAAAGCxSdl+FM248P+ijDJ+1v+qPO3Gj/aaf5Qduz/gh047x12vPjcy9a+UFue/q00/zYit5QwzPKMDSe0Ea3/aebZs0YPW/teBwaZPqmBG3awwSVnabfR3qZc4dKtT59xhopmMP8kGjbrHHaw9p4Bz/PdXdfCFHQ5qQf8KPFncVjPWrJ6b7hNBe7+1U/uZ82Ms3FrnwCWhDK5HrfLUQb00a0djdiclbPFvdQTCa/2GfA8Dw5ud+PyY45bbxxJ5DPIDeHsE0ic1YrYnLWz/w05VwYiQcZeVht2mn9SLR/bnv3tDZC0XkBYhNjuoPkEZuYGTW5hM9BE2oSDga5HojdlMdEmkb23kiO9xPg8p1y0guzLSedTzAJP5T/AEm1k15M2yxiJfK7kYUzlZePjuePu3eG55yyz3ZRdBLyzJ3L+lobcXT145zpuBbDs5jTSD9RXOhyEP877OnY2px6o+NefHkzp6l+orjRRVpx13SEHb0d56NotQUVc1dKdLmQEtOx/dHI7SKPM6YTL03REyYsuPPFSlP1k82q8PoJcHiinA4ijUZ0PudL61zeittxL9ppQ3N3ghJNrSamdYcJSn13SuFwmsh5sFEmOh6omK7u9FV0nXFutN+KSSuQc+5c1t/aCEXnp5ETs7fYyfPOLU1OOzFlH3jubMftSKznLCYT/YDhc84xXVYn0e0yBj1wOPOI8yZR6QuwndBJbC/EKiiGXzdisphMvDT1Au6IFAQ8RU6jORXiq7TeOr4Bt/1Wjt2NPYtP53J8HZb/pHJkOP9bipEcUcDeEr0RW0zImaG2HzS+rEjs4nlQtFdIJWcKh5PIR4q+QfJaoQU9Ydah3R9Xgcydy3pcG6HovBosyPGeiusXFbojPK6UsJIzIzrdaaV7XNbuFCThxXO8L+j6VejJOjbBNqAX0MPyr+tcd97SeMlblopyt0q+0x8cKOvoRU5iP1n8pybW5Vi3t5HR5lSOKcWUVnKlXtNMOyBzGWZixl7s/qPOm+OBw0rOjE50ryao1grpaF4oGf11HVrP2zF5wymKUkY9xRV0RrXnhlZep+88lz2wNkLReSVsQfY2IYZKDngMe/F8TExiAf26ix0sL6DWgD/O6crWhwCxeP4l+k/51slgvovbdCH9yUkwKujrfqXdQstKaC6L3bVamG2mKyEbtpyysvMhFaKNGH/djwCGTaDkdEXKSLT7YUnjcvH8oOOc5abjJDDKqBAq/cE8+ThiLeNTq8CaU9JyLntwbYSi8ypc0oIMfThyOSaVHPdYC+RySS6eEd+ATPgc3Ax2GYSwLnwI7Zn2+az75OU71UeRXXwahkzfcuKF8ONjTPtCKDdSTieaLsfiWbgLVVfL1biTca87Os/4Ga6Y+xxTSk7Q93OZ0soeSyKspsqnIznnlRbUcHyBbqSsORUt5rI+1kbREcBPYjz9I7cF1M2pLtcd9W2QiLd7ygse1GOuNIotysM3AJqJ3OZJ3Sh5Sv6/l6fIKXljSsvJTjM1vnCjx0Hepom2R6TvP0rdnKc/EwuvyBnUotuq8dZVU1/Pnct6XBth0flpjKPjoiDXT9E47PnPFeUPkkXtfFOas2os9GP/uFk5a4bPQT3q/J/F4d0AMA53wc0ARb2c4ign1wXtbBu62I3Gne/0bjWR/7vRTU7VD5JFxSQanff7OZhbQ6CeqZqgQn8Y7YAa9ZPh4wsppzZ+afrmT+xIitNT51W0u64SRyygLcbx3pm/bHLnso5rYxSt8ICfJLJjrNdYK4tNeieSsgrAmtOVUCZNVoIcOdno9BIyUrtgN53Ys3entZzKXb0IsYZP7CyNVcLZ4UYtBPVyfVfCvltvzSmtdX6bpyjlFhkfNdYCkCDHomPataGjZ89lrdfGOFB0XoVy4JkQGZwWZpKt71D2QssBg/pR7Mm2uf2b5dQ5PRO6HMW8Ae3atRon6Wj+WKqJH4xlDhh7Mfz+nLWIJtuyUihNSMkzGEdeaBiG74O9CYgEv51Uu+b19ey5rOXaGOOD/yO+DAAAAAAwOOCjAwAAAIDBAkUHAAAAAIMFig4AAAAABgsUHQAAAAAMFig6AAAAABgsUHQAAAAAMFig6AAAAABgsEDRAQAAAMBggaIDAAAAgMECRQcAAAAAgwWKzitRvp038Rbl1txo/6nfnt1rum/Meeu058fnXrRyX1jyiqRbvgndD72WYSD0KKfb/tNNyw91aVvlwNgLCdr2wUZKy2pLqZT9cfWZ/2p04BG2f7rdSxrHyJm2Tppt3l6vgKLzEmhB/iVa7Rb62aNwmmNa0o6u6uWtxG9RW88wkLsiB/Fsze+HVO15v4rWXdI4ZzBncNv/peVB/5FiYcnThK+CRvpj0L+cRsWX294mXHfEo3VRzBPtL8agKAeIwwrGeEm0u5o2PdFmPetBcd/QyZfVfUVT/WmJ3ljOzu6Y+iowmrrA4268L6y2FPKkNc1qx13DGJEymtF6cyrlI4YdLcctNw7ii+BHUW/cNW+BNW9dfvTtuSqd8C2v6o2x7d/+CtTbq4O3K+s369a+dTmHMp2TegNz5E2+UnZ4W3kDT5aTRdNYKj/Xb4B+dEwPCi2PoE0ebKvUvBeS6CegX7Q8U+3cNEbic55+S32LuRAWnR9nRMXXnVbBduMRbnTcH0T/mXu7mBtdpBa8pmMfJoh34nwUrbagYu7t9i4XYiPMYX98YBd6o/3fJR0WO/qvmOhnoBNPlZOF2GlueSMajDHNeUu8Ud2cIpYEQLfjXshjQ3OvcW5qgqL1syco2U82tIL15rlMJtLqGaVxjJzpKD4PLaYXuqjBTMfMwQxFZ4jcjsR6zkJ0sgp1lHVZsTlR/HXB8VUbzjzixHCcWCNOHpFsJ3Ti48aDGHz6eVvUkdWCdv/hCOpRniknG7VQC5n9iU3R2hy/OfW8gRkKeiO2mJAzQ/FR1mVFYhfPE1Q/CmkC2U+8/MET0BuMkIwxIpVRMZrdwUyfH1uanPjY+MDJZwFF5y1gJWdGdKosR4fcHgKilOfRX4WeLM/USXcUA/fvUiyZu/8oa3N5YF8TyzEPHq619CYnhzP9Y2eqzSoqs/N2Jq0FJ2g52bCSM6MT3asJ6gGFlP1CrDES+Pxoy/ZU9AhWgmvjgu7caC/NnqGFtdMYYSVnvKfi+kXG8J27YYeiM3hCJQc8hr145ugmaewjq+aUpqvKYVIGdoZl582enKGHRn9ycrntt2KSTlhzcGTVmkDJeYDQcfxEG7k5iNzUEWPnL/0XiYvx1Afn7VhdrvA3BF3GiK3kdBjMUHQGzTGp5LjHWiCXS3LxdI9Lsjj/e+zIalTQl7Tzr2mLm3QOvcrJoc6ao3ewOLLK5phScno7VprSSt6OO9Dyn6e+CDm5N6xEXD2e4MP4GLzJYGWG+JaoI9sOY+SSVnKcY606hCYLXoh+bl2pGwXRdHADpBNGLn3cADCoGwcNoSndhlsN78Yz5GRTOz71TaJSdomAsWf1/XCCUnNXn41k5GKlKfOP9YUn3M57N9JjUNBmjJRxF3dfHCqP8HkKKDovRu1EapGeKBTq8/CaZeo5aCA1ATZMjE1yCmm3ILcd8IOnk5xqNgYOOl5bZQmbi5BUm9S1Vc3CV0skTTVuIvOgjIvx1Bnd1q3XmKjcU3Nh+00LFJ0XI0/RqSbmdIeK7IzqJhHQSKhUNA24HDn55A9i01ew+3RpLadychahZnDkjc0IGHdRwk1XZM6yMO3PIbvPG9kGacbyqs8fNFCOo5ZKDpMaI5ENSji+m4Gi8wLYAzgMcYHmWQrshTadFsgnkFXDpNifRceXJYcOE8qb0E5OVdumo+k4LXaRJVB0kpTjQ4daBabRoqPHjpVe/ZwXjilsGroSa3s7NKw9dWOklLsJ7ee9D/6P+DIAAAAAwODArSsAAAAADBYoOgAAAAAYLFB0AAAAADBYoOgAAAAAYLBA0QEAAADAYIGiAwAAAIDBAkUHAAAAAIMFig4AAAAABgsUHQAAAAAMFig6AAAAABgsUHQAAAAAMFig6AAAAABgsEDRAQAAAMBggaIDAAAAgMECRQcAAAAAgwWKDgAAAAAGCxQdAAAAAAwWKDoAAAAAGCxQdAAAAAAwWKDoAAAAAGCwQNEBAAAAwGCBogMAAACAwQJFBwAAAACDBYoOAAAAAAYLFB0AAAAADBYoOgAAAAAYLFB0AAAAADBYoOgAAAAAYLBA0QEAAADAYIGiAwAAAIDBAkUHAAAAAIMFig4AAAAABgsUHQAAAAAMFig6AAAAABgsUHQAAAD8CLf9J318fFRhe9afdCdI8+OT9jf9YYTz1o77QZ91kd+R89Zpn/r2PNM2M26+7PPTTAFFBwAAwLfDCsZ4SbS73ul+53CizXpGH5976qpq8OI5Xk7pJNNT4bojWo4/KFhHb3v6FAvn7LyjqxX/qxjpCEAqOTOy2vNKu8VBtGdE2ZDtOaP15lS2Zarts2XfIs1axBcBAACA7+O6uy/E8rMRK6jDaXPnZSl4nsXpvhHfXYjV00U9JyfRVFzQiJad254sOvFssbu7LXq97xbe8xayz06zgd9p0dGauG32qjM3+qbJpIksSNfXWrUJzf++/p5bhhvtP01cz/QW27G0qpNOOxXXpBWrZ91nAADwDdyOezrQhuZT/UBzu6h5aX3sMD/dLmKmJTpcLurvOs5HWov8V7DetGc0IU9sgjMd10SLYk5ui17ochD/O+zpqJeofNnnp9nE71N0eKH+S/SfNmNxuAr17rAcRxQDpWDM1hvXlDnZBmYveV6obGllvPt9RZe/3c2oEjbHKdufSvO6o8VhSWO7AG3qJBWVMS3JNbeuLn8rpWw0p0KozKLHyIFvozrZgnZ/wq4KAADP50bHvVipFhOa6CeMPM64rEjs4sUfl/bzbjnvbZ0N6m2/FUoN0cZaWc9yBXXzB5lohdJBKo5E04mlksi1akuTk1jzxKqj9M8Wss9Os5nfp+iMCvr6KhwNb1T8R0IvCDR51cHFon5dORroqPiilbPOn+nfUjT+5uSdz05p5eXVHs5fKCImPzMY7YHcok7nf0upDZ+8+NPVF1VFH1Gx4h6zJndjpDvZZmXFBQCAn4UXuhmd6G4myoPYtat/tUDMe/+pBbD0ITlvaazn9mrOv5E0HkzFUus7xD7gH/QuqDXIVRwDWCEZ76m4inVJazTnS7xls2XfIk2fYTkjO7uAFou61ByfZOVYFDR38heD8etO91wFyqmTNuXt/kRMhx7TuVCHbDOg4PyP5Jiv66AAAPCNBAvdI/CmsXSYFYrLTMzsbKWPpb2e0V/6r7SK39khlq3tH9vQYgEkfPIhmpQXIUoul7ZC0rDIZcu+RZoxfqWiE14fHMsF3EWf42VgzgZ/kqw6xUyGSab0h01ClhlXmWtrOigAAHwjx9RC1/FYSc2jY9oXxgXhRNPlWDyL3BKKWfDl2YlvCQcSYx2LnCiUXNIKiXMEJciWfYs0U/w6RUee5YnG3hifFxlYg9cRfiHPqtNoXkgz7l56bKUcuwAA4DsZ0YTXtsOSlsFCp+YpPlZqPU+JXf9fMY+yBadSYITywpYaPs4q/S11/jE/oMlEzJn8UfDJeyOvmbNg2N/VdQWR6HZbL0OFxPiFiiiCFrLPTrOZX6bo6LPVLKvElObq7KbRCuIqBO1Rjd6VFnXS/j2H/TEcoDFGBbGrjoyPWwYAgBdhKifnyDG6dkCNHq/z8YW0did+MO5yyZ6HJxOeSCO+IDKNBRWuv8F7w+2uzqsCf9cS43sauGoY5+PqebbsW6TZyP1Xoe/P0+Je/fyB/o0EDv69enPf33t+3S2CO/zyvr53h1+m7X1XxduITxSclsxbBPc3GXLv+resk/6tATb/2Jw29vctTPygfAAA8HP4c2k57/mTsyY912rMfO/MpSYf/zuxvOrzf0sSbRpFx7XbWcks/G627FukWccvU3QYSwmQQVVYNlxUqTCKRBWSC76lFKhgC8LgpSfzVGVy081VdJi2dfLj1/3AVkyRAgCAn8coISYk52Yma9EN50YO8fkxjFub/xviy8cPQbuWMjIhtoYqsmXfIs0UH/wf8WUwWPjHBfXv7jx8VR4AAAD4XQzrejkIMVfKV1ByAAAAvB9QdIaMcSJzfiwLAAAAeB+g6AyQ8jd5zCstoOUAAAB4U+CjAwAAAIDBAosOAAAAAAYLFB0AAAAADBYoOgAAAAAYLFB0AAAAADBYoOgAAAAAYLBA0QEAAADAYIGiA0Abzlv1G0UmfO7z3iQfo3wbcyrE39J83urPH8l76PQpJ035+1RlSLxFWwM5NRO06fasP2lPKB8v1MihlNXHlrqXYKD4Y6m2359pmxk3X/b5aaaAogNAJnJgztb8gl1+Ga4IV9rRksZdJ8dRQV/qxbpeEOkuxOeLgubWezvMxHCcn2ijn4GQ3uUk4DTHyymdLDldd0TL8Qf58zPklAcrF+o3TU2bivZazzorhqPiq5SNE4Sg1HCax1+DIxZy/gF5EIGVnBlZ/Z7npoPo9xFlQ27cZrTenMq2T42RbNm3SLMW8UUAQCOxN9QL9Jt1e33rsX6LvvNmYH5Wvslev3U5683478Yz5JRI08jBFhTklIeWR/D261jffxD1luzUG68rGdbHAyXmbeKekGT7BX39ehebNvd5C9lnp9nAC1h0+O3aH/Qp1EOzE1JaXWWu4s8cIib/II5FZZbUIdyCyfTkY89MF023lSkvk8Y6qXaKm1ZTn/kmPyu0UocBnY+0FvvCwjaxMJcLHcT/DvujuxPpjJDllreXG5rbb+6YruiOt8838ww53S5yXB1EGo1ATlncjnshD6+PC24XNS+tjz3NT2JeVcNpTrEX4Zy3M9FfNnTCa3LyGU0ibXmmo2jn0Gp2oYsaeHTUAy9f9vlpNvEyR1eH5ZjGFzFJsJlRVODvp7SX0VWoboflv2oBZ4XgL9F/2ozFQcUZR5QStdDP1qIj2/En26jZaz0TCoBlpjPp2nGl0jQ7Wya3O502bMpraUqzyarTiOaF0IPFsAzmgNuR9kLwi92fqgNKxck3+fH3OR7ef9WWM4840boTa8RJxXw7oRO360EMPv38IWKyBNk8RU6jOamht3U2NLf9VoxGXkMhqXbc6Cg7+YQm+gkjjzPEGiB28eKPSy8bB7WoLmj3JyIjfWS1Oa0w1tqgFX8HucEQI88deGIN2tLkxEeHB95r8MN82Wen2cwL+ejYnfFAh6n9xu0zXcr1vqAvb8c0Kv6TPg3+jktNRCLdq9uR+Sw3us4vdnS9V3FH80J8m9tdZy4amXcHm9MXFVYBpivlU7HednQ6zKzTqFjJM39/t2MGs72LjQ1w8/2snSmoRfps7Au6CrmpAWv10Qc4/1sKuW1oZXcw0Jl+5DSi4j81sZa+CWKRHC/FhC02EtgzPA4vdDMSmzLTmL1sHM70T8po5czXCrEJVloO5NcSNUeJpqtT8FkhGe+puIq1Ums05TrqkS37Fmn6vI6i4zleLia2vpeJswvQmmO0k8cJTGTaWfRLJ5AyufFEKK0tfe3qbZw6TWmuNB0xTA3t6wkew148e21yrUjDmtMPvcpJzgXGEZOtumILBctoLwQLXU+UG92INQdHVt3gMSUdtxc7ihnJJLZC0jDwsmXfIs0Yv/LWFTe263MyJlbcXYTSETx7jEtjgt139Xl1EqrOH95ZWsdX+qjD166VNUpMyv8slQim9oe5JBdP97ikC7Dm9EffclLjc0z7Qig38ij4RNPlWDzrwT/vjTmmFjrvaKM9NdYcvaHAkVVLjBWTFcTU5uGSVkicIyhBtuxbpJni1yk68ixPNPamvDrKQV/HfTKTSToTpQR1m0Rb1Un7C5jjK2llimnXF+V8SXxlTytP5jofNjHtUbJf0zJYPONnzq3Rk2/KaRLk8RQ5Cdn8FeOTLTjGustjfcVXYnkz8bfjkfXbMqIJd/LDkpbBQqccUMUKFl9IM6nb1CnLPE+N9sZSLLr8BfGtmfwbv6fjwBdwZAOxv2tEQZxMxMZatN4yVEiMG4U6pGkh++w0m/llis6NpGN2ndmsJHbM8xgjKaGIM7ApV6dFqk2dGH1MJuuldi2x34eQTpnS58hWnrqZ/QDrl8pfK2hr4zwckQHDSqycSBs81ZU1J+E0CbLpJqfqdmJUTGbTAHpjKifniCKiHVCjVmc+vpByarKiaWtOYk5N/d6OdISVCzn/DWtPCbe7Oq8K/F1LjMO+54LC65vaYFTPs2XfIs1GhIB/GO9OvPd7F9fd4i5qdFd/6rjl34z+HQQO/r16c9/fe85pOnf4vTzTxPLXd/29Z/m0rJNEfb7Zcbnjv/ugyoTfhOgTty8yTb/nYMmxThaJ36VIo9NN5vvetJaT/v0OGWIyMPKJjvu6eQNyqiOco3R7JcaBkmtTm1fxsoeTBnNmhETfjxJZR8OxqMiWfYs06/hlig5jLx4c1Gey4aITilEkqhAMkkhj1mEmuDI8PJG1rZNVhuRo9tO0QtsZAJTYk21OWzbLycRpGrg18uQAmTq0k1PVtulo8fYP40NObfDn0to5OGvR1e3fYU5WZYGiY+PLxw9Bdy5lZEK6PbNl3yLNFB/8H/FlMCj4BwTZmTlynmrOWvn3deCsAwAAYOD8yltXoAHtjxD1GdIOXgAAAMA7AEVniJif6I44Ymf92BMAAAAwEHB0NVjM8ZX+syRxPRAAAAAYIFB0AAAAADBYcHQFAAAAgMECRQcAAAAAgwWKDgAAAAAGCxQdAAAAAAwWKDoAAAAAGCxQdAAAAAAwWL5J0Wl4O3AL1NugX+wV+vqtup/1r9QFAAAAwDczfIsOv9upBwULAInuT2X43NPD6q1WlKt004r8bf9pxRMBHTvOE+QUtP3HJ9XubZ7RVwbGM/pzs5yqjXcqYFhZ+P24tt/7bZuOmy/7/DRTfJOiM6WVelM64T2S4LciB+ZszS+gln35fr/SjpY0fsDCKNMcL2lapnmn6+5Ms8hgZmumiEq7q4l7os16hgXU41lyGi+ndNIyUnIiWo5jiyL/KrmYkGdnS1YifBU00jHAc/pznpyq9cgPpw1/viG8IUfDSs6MrPYUY2lxEO0ZUTbkhm1Ga35htI6fGiPZsm+RZi3ii7+K1q/SP234l5/D18n3iX6NfPI182AAnO5iDgxl/JDsE2ner/fdQvRzu9PqfIJ+/B39+1fxnXJSzx05CeQctdgJKYIkT+nP7eQUgHk8D91Oef1ez2X28xayz06zgf8Zs1DavyT2ud6xWOak6PcDk3y9yUn538RCbCfmlcFR7yxTl9jZMeuZFVeEsLyZdWL8egnVNHilVFuCtkq1eV1b+J/5Jj8rtFKHAZ2PtKYFFXNvT365SNkf9kchhZbcLlJe04m/zx/RhHeU50uZ5u24F/mEO83bRclxfYQ8JU+U00Gk0cyZjmLK2axgvanjKf25lZxC1AuPI30HuJiXRjuofr8o5l6/v9BFDTw66oGXL/v8NJsoj66SnUN3ngpePMe0pB0JbUqZk5QtKVw8RwV9leYmocMlUQv1bM0vnKzSlN9YcD7eSygXqgyXlRV3PbMUA8s0qWyRlhlbha/Cbrr8Oqmjhj0VtknalLUrrOT8JfrPpCcCt9dB5F/VaUTzgnNZUzAH3I60F4Jf7P5U7RQ1+alSih0L3XGG2IozjzjRurZOIvvCdkInbteDGHz6eVvOl3C0TiZ2mjc6SgFPaCL/Vkjz72WlzO2WUvTOPEVOozmpobd1Nmq3/VaMRjG32DO2VrQmtqCAx5P6cxs5BahFVWio5CwNICTQCQSy33ubNrkGbWly4vXxwHsNfpgv++w0mxGKjsiQO0eJsgL41gyTmeo0QiGxz5uFQvMfTyLrY9gAOcQWapHmiisdm5gOB+nTUK7VuoN32q0JsuskGvnvUuj8u//6HQysEHrn96PiP1LzclX7UbESpeQiecqX1JDdnYh5tvtTDW7z/a47HlAhfQH2BV2F3NSAPVNEX6nH9NvlX9fSed7SOHztvANPDDMSSqwZBA8oWkOmFzmJkVn8pybW0jfByEhsJOw9g9qVCkWLfAttewfKd6Kf/pwvJx+1BrjzJYijLF8NiiMrJNIg8EWF1mhiGzomW/Yt0vQRio5nJtdaVKk0iEWRF0y1Q9Ha2GZeKSSakUwkYm14Cn06i7Wok26LbzVtOjubKc2VpmMplKb82Il8F/bi+ViTi4n5ix0geWK2FsXjvNYCGkwMIEp/chJI67RxxBQymollMWkZXdPMs9AaGePEOKTX/txKToYz/ZPKEObQJnhMSW+QxY6SOqGtkDS0Z7bsW6QZQx5d2WZy3pEsdjva6POvcociExdxWJVj72hnt6I6VGfKne0/x3qy5SQjCki/5NfJnCE+A+5Abv5jim3qp394x2IpX9oa5mvXo3mhdjb/qjLnmXBBHZfk4ukel7Rhao5gTRCD/sKd0jPvMsfUxBCJ+870LSc1Pse0L8SiKeV0oikfbUctNZ51WDBdnaLW2Hen7/7cTk4KWHMyKS3NYf8uuaQVEt8XMVv2LdJMIjqD9nZe3HdX9mY2/1fe52JnaXk3R26DZCLTkXnoBw7aK94LMe936YUd3Lqq8cJu9OLPr1OyDg9666s6+WVM1cktrysfC11vN6TaHzShZC/aMCWPmAw6E94SMX0k7KeZN0rehKfIKTm+ddvbaZZzqf675Bn95PfylP7cRk4ldZ+BknI9Sdx4NjexIn3fXzezZd8izSaUM/JkIvRZweVIQm+i+Ug5vvLx1ZF3ltOJ1t70MVdXX5wU2oFPFFpr4Sq4DsPPIr9O6igrdIAyZ5bduJE0FNWZAh20U7IsrzK5hl7p3KRrmaboB1abdjP7AWMhE03qt7XxL4vIgGHTrLTQtTiziO0wp/LMMmKN00fNsNIpusmpup0YFZM8ss5EzqUxJ0llOU71k3ejU3/m4wspp4R1po2cNKWVG7fk0nC7q/MqsUZ7F4MMxhF8wfqDeqQwzsfV82zZt0izEbEACrQmFdWmPA3ZPG+pAddqYI1Wl4rWFp0cjT27TmE+sjybjcyjm0VHp+m0jSWPmjptdlzuuIYdbyfwCGEfbtqlW3LMkoXpC/GxEMpUp58zcN6I1nKyrZ+xtkzsLM3O1B/3sbGH8RjStj8rucbbXNJSTmV+jfP+G5No0yg6rt3OqXU/W/Yt0qxDKzqxCbZu0q0+c4LXYeyOGYbYRBALORNGw0RmGtEKYafPq5OflkpHPQvTzMUvnxKirGuiTmV7JSaFME0rJL8Dmgj6dENbNsmpc3o6dO9zw6Zdu1ZjJR0tPp5S8X05YTGN06o/Zy26+XIyfaRhyL01QT/2QtB2pYxM8NfqimzZt0gzxQf/R3z5R5G3I5YHnou8K4BsUp6RvPrt/5YOqIF/l4idmSPtJn/Sey36Sk+3HAAAAIAXpvzBwJ9Dn7dFr4z7v/EDstD+CBS7sWb8sQAAAIA34AUUHe0MTJHf4Dn/U1esn37FfGCYn+iOOFhn/dgTAAAAMBBe4uiKMcdXPuFxFsjDHF/pP0twDAgAAOB9eBlFBwAAAACgb17g6AoAAAAA4DlA0QEAAADAYIGiAwAAAIDBAkUHAAAAAIMFig4AAAAABgsUHQAAAAAMFig6AAAAABgsUHQAAAAAMFig6AAAAABgsEDRAQAAAMBAIfo/g0U7TQxNxAcAAAAASUVORK5CYII="
    }
   },
   "cell_type": "markdown",
   "id": "32c71b1d",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 396,
   "id": "3a34aa0e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "pickle.dump(gb_model, open('best_model.pkl', 'wb'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ae3190af",
   "metadata": {},
   "source": [
    "## RFECV"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd693189",
   "metadata": {},
   "source": [
    "Best model params:  \n",
    "The best parameters are: {'model__learning_rate': 0.1, 'model__max_features': 5, 'model__n_estimators': 100, 'model__subsample': 1}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 403,
   "id": "06841cec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-3 {color: black;background-color: white;}#sk-container-id-3 pre{padding: 0;}#sk-container-id-3 div.sk-toggleable {background-color: white;}#sk-container-id-3 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-3 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-3 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-3 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-3 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-3 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-3 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-3 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-3 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-3 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-3 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-3 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-3 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-3 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-3 div.sk-item {position: relative;z-index: 1;}#sk-container-id-3 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-3 div.sk-item::before, #sk-container-id-3 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-3 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-3 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-3 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-3 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-3 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-3 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-3 div.sk-label-container {text-align: center;}#sk-container-id-3 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-3 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-3\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RFECV(cv=StratifiedKFold(n_splits=5, random_state=None, shuffle=False),\n",
       "      estimator=GradientBoostingClassifier(max_features=5, random_state=42,\n",
       "                                           subsample=1),\n",
       "      scoring=&#x27;roc_auc&#x27;)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RFECV</label><div class=\"sk-toggleable__content\"><pre>RFECV(cv=StratifiedKFold(n_splits=5, random_state=None, shuffle=False),\n",
       "      estimator=GradientBoostingClassifier(max_features=5, random_state=42,\n",
       "                                           subsample=1),\n",
       "      scoring=&#x27;roc_auc&#x27;)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: GradientBoostingClassifier</label><div class=\"sk-toggleable__content\"><pre>GradientBoostingClassifier(max_features=5, random_state=42, subsample=1)</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-5\" type=\"checkbox\" ><label for=\"sk-estimator-id-5\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">GradientBoostingClassifier</label><div class=\"sk-toggleable__content\"><pre>GradientBoostingClassifier(max_features=5, random_state=42, subsample=1)</pre></div></div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "RFECV(cv=StratifiedKFold(n_splits=5, random_state=None, shuffle=False),\n",
       "      estimator=GradientBoostingClassifier(max_features=5, random_state=42,\n",
       "                                           subsample=1),\n",
       "      scoring='roc_auc')"
      ]
     },
     "execution_count": 403,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use all data\n",
    "X = eng_data[features]\n",
    "y = eng_data.Exited\n",
    "# Scale data\n",
    "cat_cols = ['HasCrCard', 'IsActiveMember', 'Complain', 'Card Type', 'Female', 'Male', \n",
    "            'France','Germany','Spain','zero_balance','high_balance_low_salary','high_salary_low_balance']\n",
    "scale_num_cols = custom_scaler(variables=cat_cols, excludeComplain=True)\n",
    "X_scaled = scale_num_cols.fit_transform(X)\n",
    "# Build rfe\n",
    "estimator = GradientBoostingClassifier(random_state=42, learning_rate= 0.1, \n",
    "                                       max_features=5, n_estimators=100, subsample=1)\n",
    "rfecv = RFECV(estimator=estimator, cv=StratifiedKFold(n_splits=5), scoring='roc_auc')\n",
    "rfecv.fit(X_scaled, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 406,
   "id": "0d2cf3f4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['CreditScore', 'Age', 'Tenure', 'Balance', 'NumOfProducts',\n",
       "       'EstimatedSalary', 'Satisfaction Score', 'Point Earned',\n",
       "       'tenure_balance_ratio', 'salary_balance_ratio', 'IsActiveMember',\n",
       "       'Female', 'Male', 'France', 'Germany', 'Spain', 'zero_balance',\n",
       "       'high_balance_low_salary'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 406,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "selected_features = X_scaled.columns[rfecv.support_]\n",
    "selected_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 409,
   "id": "4600c76d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "17"
      ]
     },
     "execution_count": 409,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(rfecv.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 416,
   "id": "58437fda",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "plt.grid(True)\n",
    "plt.xlabel(\"Number of features selected\")\n",
    "plt.ylabel(\"Cross-validation score (roc_auc)\")\n",
    "plt.plot(range(1, len(rfecv.cv_results_['mean_test_score']) + 1), rfecv.cv_results_['mean_test_score'],\n",
    "        marker='o', color='#3282b8', markerfacecolor = 'None', markeredgecolor = '#3282b8', markeredgewidth = 1.2)\n",
    "plt.plot(18, rfecv.cv_results_['mean_test_score'][17], marker='o', markerfacecolor = '#3282b8', markeredgewidth = 1.2)\n",
    "plt.plot(5, rfecv.cv_results_['mean_test_score'][4], marker='o', markerfacecolor = 'r', markeredgewidth = 1.2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd3654e0",
   "metadata": {},
   "source": [
    "The roc_auc score peaks with 18 features selected. The score stops increasing after 5 features and this is the optimal number of features."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8da87b21",
   "metadata": {},
   "source": [
    "### Features importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 455,
   "id": "2d1efd67",
   "metadata": {},
   "outputs": [],
   "source": [
    "f1 = features.copy()\n",
    "f1.remove('Complain')\n",
    "X_train_sc = scale_num_cols.fit_transform(X_train[f1])\n",
    "estimator.fit(X_train_sc, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 456,
   "id": "f0bc55af",
   "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>CreditScore</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <th>Point Earned</th>\n",
       "      <th>female_male_ratio</th>\n",
       "      <th>tenure_balance_ratio</th>\n",
       "      <th>...</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>Card Type</th>\n",
       "      <th>Female</th>\n",
       "      <th>Male</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Spain</th>\n",
       "      <th>zero_balance</th>\n",
       "      <th>high_balance_low_salary</th>\n",
       "      <th>high_salary_low_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9254</th>\n",
       "      <td>0.356500</td>\n",
       "      <td>-0.655786</td>\n",
       "      <td>0.345680</td>\n",
       "      <td>-1.218471</td>\n",
       "      <td>0.808436</td>\n",
       "      <td>1.367670</td>\n",
       "      <td>-0.720010</td>\n",
       "      <td>-0.430193</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>-0.860772</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1561</th>\n",
       "      <td>-0.203898</td>\n",
       "      <td>0.294938</td>\n",
       "      <td>-0.348369</td>\n",
       "      <td>0.696838</td>\n",
       "      <td>0.808436</td>\n",
       "      <td>1.661254</td>\n",
       "      <td>0.704342</td>\n",
       "      <td>1.565908</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>0.151162</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1670</th>\n",
       "      <td>-0.961472</td>\n",
       "      <td>-1.416365</td>\n",
       "      <td>-0.695393</td>\n",
       "      <td>0.618629</td>\n",
       "      <td>-0.916688</td>\n",
       "      <td>-0.252807</td>\n",
       "      <td>0.704342</td>\n",
       "      <td>-1.243749</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>-0.069512</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6087</th>\n",
       "      <td>-0.940717</td>\n",
       "      <td>-1.131148</td>\n",
       "      <td>1.386753</td>\n",
       "      <td>0.953212</td>\n",
       "      <td>-0.916688</td>\n",
       "      <td>0.915393</td>\n",
       "      <td>-0.720010</td>\n",
       "      <td>-0.176791</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>1.147289</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6669</th>\n",
       "      <td>-1.397337</td>\n",
       "      <td>1.625953</td>\n",
       "      <td>1.386753</td>\n",
       "      <td>1.057449</td>\n",
       "      <td>-0.916688</td>\n",
       "      <td>-1.059600</td>\n",
       "      <td>-0.007834</td>\n",
       "      <td>0.534515</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>1.055320</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5734</th>\n",
       "      <td>1.207474</td>\n",
       "      <td>1.435808</td>\n",
       "      <td>1.039728</td>\n",
       "      <td>-0.102301</td>\n",
       "      <td>-0.916688</td>\n",
       "      <td>-0.539860</td>\n",
       "      <td>-0.007834</td>\n",
       "      <td>-0.167899</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>2.612112</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5191</th>\n",
       "      <td>0.314989</td>\n",
       "      <td>1.816097</td>\n",
       "      <td>-1.389442</td>\n",
       "      <td>-1.218471</td>\n",
       "      <td>-0.916688</td>\n",
       "      <td>-1.733882</td>\n",
       "      <td>0.704342</td>\n",
       "      <td>0.454493</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>-0.860772</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5390</th>\n",
       "      <td>0.865009</td>\n",
       "      <td>-0.085351</td>\n",
       "      <td>-1.389442</td>\n",
       "      <td>-1.218471</td>\n",
       "      <td>2.533560</td>\n",
       "      <td>-0.142765</td>\n",
       "      <td>1.416518</td>\n",
       "      <td>0.316678</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>-0.860772</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>860</th>\n",
       "      <td>0.159323</td>\n",
       "      <td>0.390011</td>\n",
       "      <td>1.039728</td>\n",
       "      <td>1.827259</td>\n",
       "      <td>-0.916688</td>\n",
       "      <td>-0.050826</td>\n",
       "      <td>-0.007834</td>\n",
       "      <td>0.325569</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>0.411938</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7270</th>\n",
       "      <td>0.470655</td>\n",
       "      <td>1.150590</td>\n",
       "      <td>-1.389442</td>\n",
       "      <td>1.149720</td>\n",
       "      <td>-0.916688</td>\n",
       "      <td>-0.814568</td>\n",
       "      <td>0.704342</td>\n",
       "      <td>0.143297</td>\n",
       "      <td>-3.330669e-16</td>\n",
       "      <td>-0.656168</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8000 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      CreditScore       Age    Tenure   Balance  NumOfProducts  \\\n",
       "9254     0.356500 -0.655786  0.345680 -1.218471       0.808436   \n",
       "1561    -0.203898  0.294938 -0.348369  0.696838       0.808436   \n",
       "1670    -0.961472 -1.416365 -0.695393  0.618629      -0.916688   \n",
       "6087    -0.940717 -1.131148  1.386753  0.953212      -0.916688   \n",
       "6669    -1.397337  1.625953  1.386753  1.057449      -0.916688   \n",
       "...           ...       ...       ...       ...            ...   \n",
       "5734     1.207474  1.435808  1.039728 -0.102301      -0.916688   \n",
       "5191     0.314989  1.816097 -1.389442 -1.218471      -0.916688   \n",
       "5390     0.865009 -0.085351 -1.389442 -1.218471       2.533560   \n",
       "860      0.159323  0.390011  1.039728  1.827259      -0.916688   \n",
       "7270     0.470655  1.150590 -1.389442  1.149720      -0.916688   \n",
       "\n",
       "      EstimatedSalary  Satisfaction Score  Point Earned  female_male_ratio  \\\n",
       "9254         1.367670           -0.720010     -0.430193      -3.330669e-16   \n",
       "1561         1.661254            0.704342      1.565908      -3.330669e-16   \n",
       "1670        -0.252807            0.704342     -1.243749      -3.330669e-16   \n",
       "6087         0.915393           -0.720010     -0.176791      -3.330669e-16   \n",
       "6669        -1.059600           -0.007834      0.534515      -3.330669e-16   \n",
       "...               ...                 ...           ...                ...   \n",
       "5734        -0.539860           -0.007834     -0.167899      -3.330669e-16   \n",
       "5191        -1.733882            0.704342      0.454493      -3.330669e-16   \n",
       "5390        -0.142765            1.416518      0.316678      -3.330669e-16   \n",
       "860         -0.050826           -0.007834      0.325569      -3.330669e-16   \n",
       "7270        -0.814568            0.704342      0.143297      -3.330669e-16   \n",
       "\n",
       "      tenure_balance_ratio  ...  IsActiveMember  Card Type  Female  Male  \\\n",
       "9254             -0.860772  ...             1.0        0.0     0.0   1.0   \n",
       "1561              0.151162  ...             1.0        2.0     0.0   1.0   \n",
       "1670             -0.069512  ...             0.0        3.0     0.0   1.0   \n",
       "6087              1.147289  ...             0.0        3.0     1.0   0.0   \n",
       "6669              1.055320  ...             0.0        2.0     0.0   1.0   \n",
       "...                    ...  ...             ...        ...     ...   ...   \n",
       "5734              2.612112  ...             1.0        3.0     0.0   1.0   \n",
       "5191             -0.860772  ...             1.0        2.0     1.0   0.0   \n",
       "5390             -0.860772  ...             0.0        3.0     1.0   0.0   \n",
       "860               0.411938  ...             0.0        0.0     0.0   1.0   \n",
       "7270             -0.656168  ...             1.0        0.0     0.0   1.0   \n",
       "\n",
       "      France  Germany  Spain  zero_balance  high_balance_low_salary  \\\n",
       "9254     1.0      0.0    0.0           1.0                      1.0   \n",
       "1561     0.0      1.0    0.0           0.0                      0.0   \n",
       "1670     0.0      0.0    1.0           0.0                      0.0   \n",
       "6087     1.0      0.0    0.0           0.0                      0.0   \n",
       "6669     1.0      0.0    0.0           0.0                      0.0   \n",
       "...      ...      ...    ...           ...                      ...   \n",
       "5734     1.0      0.0    0.0           0.0                      0.0   \n",
       "5191     1.0      0.0    0.0           1.0                      1.0   \n",
       "5390     1.0      0.0    0.0           1.0                      1.0   \n",
       "860      1.0      0.0    0.0           0.0                      0.0   \n",
       "7270     0.0      1.0    0.0           0.0                      0.0   \n",
       "\n",
       "      high_salary_low_balance  \n",
       "9254                      0.0  \n",
       "1561                      1.0  \n",
       "1670                      0.0  \n",
       "6087                      0.0  \n",
       "6669                      0.0  \n",
       "...                       ...  \n",
       "5734                      0.0  \n",
       "5191                      0.0  \n",
       "5390                      0.0  \n",
       "860                       0.0  \n",
       "7270                      0.0  \n",
       "\n",
       "[8000 rows x 22 columns]"
      ]
     },
     "execution_count": 456,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_sc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 431,
   "id": "2bfaaf59",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['female_male_ratio', 'HasCrCard', 'Card Type', 'Satisfaction Score',\n",
       "       'high_salary_low_balance', 'high_balance_low_salary', 'Tenure', 'Spain',\n",
       "       'France', 'Female', 'zero_balance', 'EstimatedSalary', 'Point Earned',\n",
       "       'Male', 'tenure_balance_ratio', 'CreditScore', 'salary_balance_ratio',\n",
       "       'Balance', 'Germany', 'IsActiveMember', 'NumOfProducts', 'Age'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 431,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sort_indices = estimator.feature_importances_.argsort()\n",
    "X_train_sc.columns[sort_indices]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 432,
   "id": "b8dd09c0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "plt.title(\"Feature Importance\")\n",
    "plt.barh(range(X_train_sc.shape[1]), estimator.feature_importances_[sort_indices])\n",
    "plt.yticks(range(X_train_sc.shape[1]), X_train_sc.columns[sort_indices], rotation=0)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e78b919",
   "metadata": {},
   "source": [
    "From the chart it can be seen that the top 3 features are age, number of products with the bank, and active member status. However, age and number of products are the most influential features in determining churn due to its contribution value.\n",
    "\n",
    "Overall, the bank institution should focus on building better experience for their customer. This may be solved by resolving how the bank handles complains received as complain is the main driver. Furthermore, age is also important in determining churn, the further exploration of which age group churned in the past will be conducted. Lastly, number of products that the customer has with the bank is also crucial. Increasing bank products as well as its quality will prevent customer from churning."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71c58782",
   "metadata": {},
   "source": [
    "### Exploring Age group"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 437,
   "id": "f3f75665",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    10000.000000\n",
       "mean        38.921800\n",
       "std         10.487806\n",
       "min         18.000000\n",
       "25%         32.000000\n",
       "50%         37.000000\n",
       "75%         44.000000\n",
       "max         92.000000\n",
       "Name: Age, dtype: float64"
      ]
     },
     "execution_count": 437,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['Age'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 436,
   "id": "e5a5eadf",
   "metadata": {},
   "outputs": [],
   "source": [
    "group = [18,20,30,40,50,60,70,93]\n",
    "labels = ['<20', '20s', '30s','40s','50s','60s','>70']\n",
    "data['Age_group']=pd.cut(data['Age'], bins=group, labels=labels, right=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 438,
   "id": "3b8c14da",
   "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>CustomerId</th>\n",
       "      <th>Surname</th>\n",
       "      <th>CreditScore</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>HasCrCard</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>...</th>\n",
       "      <th>Complain</th>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <th>Card Type</th>\n",
       "      <th>Point Earned</th>\n",
       "      <th>Female</th>\n",
       "      <th>Male</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Spain</th>\n",
       "      <th>Age_group</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>15634602</td>\n",
       "      <td>Hargrave</td>\n",
       "      <td>619</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>464</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15647311</td>\n",
       "      <td>Hill</td>\n",
       "      <td>608</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>456</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>15619304</td>\n",
       "      <td>Onio</td>\n",
       "      <td>502</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>377</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>15701354</td>\n",
       "      <td>Boni</td>\n",
       "      <td>699</td>\n",
       "      <td>39</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>93826.63</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>GOLD</td>\n",
       "      <td>350</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>30s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15737888</td>\n",
       "      <td>Mitchell</td>\n",
       "      <td>850</td>\n",
       "      <td>43</td>\n",
       "      <td>2</td>\n",
       "      <td>125510.82</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>79084.10</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>GOLD</td>\n",
       "      <td>425</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>15574012</td>\n",
       "      <td>Chu</td>\n",
       "      <td>645</td>\n",
       "      <td>44</td>\n",
       "      <td>8</td>\n",
       "      <td>113755.78</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>149756.71</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>484</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>15592531</td>\n",
       "      <td>Bartlett</td>\n",
       "      <td>822</td>\n",
       "      <td>50</td>\n",
       "      <td>7</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10062.80</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>SILVER</td>\n",
       "      <td>206</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>50s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>15656148</td>\n",
       "      <td>Obinna</td>\n",
       "      <td>376</td>\n",
       "      <td>29</td>\n",
       "      <td>4</td>\n",
       "      <td>115046.74</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>119346.88</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>DIAMOND</td>\n",
       "      <td>282</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>15792365</td>\n",
       "      <td>He</td>\n",
       "      <td>501</td>\n",
       "      <td>44</td>\n",
       "      <td>4</td>\n",
       "      <td>142051.07</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>74940.50</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>GOLD</td>\n",
       "      <td>251</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40s</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>15592389</td>\n",
       "      <td>H?</td>\n",
       "      <td>684</td>\n",
       "      <td>27</td>\n",
       "      <td>2</td>\n",
       "      <td>134603.88</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>71725.73</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>GOLD</td>\n",
       "      <td>342</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20s</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustomerId   Surname  CreditScore  Age  Tenure    Balance  NumOfProducts  \\\n",
       "0    15634602  Hargrave          619   42       2       0.00              1   \n",
       "1    15647311      Hill          608   41       1   83807.86              1   \n",
       "2    15619304      Onio          502   42       8  159660.80              3   \n",
       "3    15701354      Boni          699   39       1       0.00              2   \n",
       "4    15737888  Mitchell          850   43       2  125510.82              1   \n",
       "5    15574012       Chu          645   44       8  113755.78              2   \n",
       "6    15592531  Bartlett          822   50       7       0.00              2   \n",
       "7    15656148    Obinna          376   29       4  115046.74              4   \n",
       "8    15792365        He          501   44       4  142051.07              2   \n",
       "9    15592389        H?          684   27       2  134603.88              1   \n",
       "\n",
       "   HasCrCard  IsActiveMember  EstimatedSalary  ...  Complain  \\\n",
       "0          1               1        101348.88  ...         1   \n",
       "1          0               1        112542.58  ...         1   \n",
       "2          1               0        113931.57  ...         1   \n",
       "3          0               0         93826.63  ...         0   \n",
       "4          1               1         79084.10  ...         0   \n",
       "5          1               0        149756.71  ...         1   \n",
       "6          1               1         10062.80  ...         0   \n",
       "7          1               0        119346.88  ...         1   \n",
       "8          0               1         74940.50  ...         0   \n",
       "9          1               1         71725.73  ...         0   \n",
       "\n",
       "   Satisfaction Score  Card Type Point Earned  Female  Male  France  Germany  \\\n",
       "0                   2    DIAMOND          464       1     0       1        0   \n",
       "1                   3    DIAMOND          456       1     0       0        0   \n",
       "2                   3    DIAMOND          377       1     0       1        0   \n",
       "3                   5       GOLD          350       1     0       1        0   \n",
       "4                   5       GOLD          425       1     0       0        0   \n",
       "5                   5    DIAMOND          484       0     1       0        0   \n",
       "6                   2     SILVER          206       0     1       1        0   \n",
       "7                   2    DIAMOND          282       1     0       0        1   \n",
       "8                   3       GOLD          251       0     1       1        0   \n",
       "9                   3       GOLD          342       0     1       1        0   \n",
       "\n",
       "   Spain  Age_group  \n",
       "0      0        40s  \n",
       "1      1        40s  \n",
       "2      0        40s  \n",
       "3      0        30s  \n",
       "4      1        40s  \n",
       "5      1        40s  \n",
       "6      0        50s  \n",
       "7      0        20s  \n",
       "8      0        40s  \n",
       "9      0        20s  \n",
       "\n",
       "[10 rows x 21 columns]"
      ]
     },
     "execution_count": 438,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 439,
   "id": "9e6b097d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='Age_group', ylabel='count'>"
      ]
     },
     "execution_count": 439,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x=data['Age_group'], hue=data['Exited'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52195126",
   "metadata": {},
   "source": [
    "Most of the customers who churn are located in the 40s age group."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7878d7fc",
   "metadata": {},
   "source": [
    "## Resampling Method (Equal samples for each class target)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d2071fca",
   "metadata": {},
   "source": [
    "### Using resample technique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 506,
   "id": "39686e8c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.utils import resample\n",
    "\n",
    "# combined features and target data\n",
    "train_data = pd.concat([X_train, y_train], axis=1)\n",
    "major_class = train_data.loc[train_data['Exited']==0]\n",
    "minor_class = train_data.loc[train_data['Exited']==1]\n",
    "# resample until minor class reaches half of major class\n",
    "minor_upsampled = resample(minor_class, replace=True, n_samples=int(len(major_class)/2),\n",
    "                          random_state=42)\n",
    "upsampled_df = pd.concat([major_class, minor_upsampled])\n",
    "# define feature and target\n",
    "X_train = upsampled_df.drop('Exited', axis=1)\n",
    "y_train = upsampled_df['Exited']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 507,
   "id": "4ba8adda",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    6355\n",
       "1.0    3177\n",
       "Name: Exited, dtype: int64"
      ]
     },
     "execution_count": 507,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_re.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 508,
   "id": "45c6347c",
   "metadata": {
    "scrolled": false
   },
   "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>CreditScore</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>HasCrCard</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>Complain</th>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <th>...</th>\n",
       "      <th>Male</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Spain</th>\n",
       "      <th>zero_balance</th>\n",
       "      <th>female_male_ratio</th>\n",
       "      <th>tenure_balance_ratio</th>\n",
       "      <th>salary_balance_ratio</th>\n",
       "      <th>high_balance_low_salary</th>\n",
       "      <th>high_salary_low_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9254</th>\n",
       "      <td>0.010431</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.010368</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.015515</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1561</th>\n",
       "      <td>0.009610</td>\n",
       "      <td>42.0</td>\n",
       "      <td>0.006912</td>\n",
       "      <td>0.012119</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.016977</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.570320</td>\n",
       "      <td>1.400882</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8829</th>\n",
       "      <td>0.010385</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0.001728</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.006563</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7945</th>\n",
       "      <td>0.011541</td>\n",
       "      <td>45.0</td>\n",
       "      <td>0.013824</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.008598</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3508</th>\n",
       "      <td>0.006477</td>\n",
       "      <td>34.0</td>\n",
       "      <td>0.005184</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.005304</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6564</th>\n",
       "      <td>0.009914</td>\n",
       "      <td>47.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.012826</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.003361</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.262062</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1685</th>\n",
       "      <td>0.009321</td>\n",
       "      <td>20.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.011889</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.009837</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.827415</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>916</th>\n",
       "      <td>0.007222</td>\n",
       "      <td>39.0</td>\n",
       "      <td>0.010368</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.004938</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>0.012681</td>\n",
       "      <td>49.0</td>\n",
       "      <td>0.003456</td>\n",
       "      <td>0.013312</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.016838</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.259616</td>\n",
       "      <td>1.264897</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1003</th>\n",
       "      <td>0.009823</td>\n",
       "      <td>42.0</td>\n",
       "      <td>0.005184</td>\n",
       "      <td>0.017745</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.005815</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.292123</td>\n",
       "      <td>0.327687</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9532 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      CreditScore   Age    Tenure   Balance  NumOfProducts  HasCrCard  \\\n",
       "9254     0.010431  32.0  0.010368  0.000000            2.0        1.0   \n",
       "1561     0.009610  42.0  0.006912  0.012119            2.0        1.0   \n",
       "8829     0.010385  40.0  0.001728  0.000000            2.0        0.0   \n",
       "7945     0.011541  45.0  0.013824  0.000000            2.0        1.0   \n",
       "3508     0.006477  34.0  0.005184  0.000000            2.0        1.0   \n",
       "...           ...   ...       ...       ...            ...        ...   \n",
       "6564     0.009914  47.0  0.000000  0.012826            2.0        1.0   \n",
       "1685     0.009321  20.0  0.000000  0.011889            1.0        0.0   \n",
       "916      0.007222  39.0  0.010368  0.000000            1.0        1.0   \n",
       "43       0.012681  49.0  0.003456  0.013312            1.0        0.0   \n",
       "1003     0.009823  42.0  0.005184  0.017745            2.0        0.0   \n",
       "\n",
       "      IsActiveMember  EstimatedSalary  Complain  Satisfaction Score  ...  \\\n",
       "9254             1.0         0.015515       0.0                 2.0  ...   \n",
       "1561             1.0         0.016977       0.0                 4.0  ...   \n",
       "8829             0.0         0.006563       0.0                 3.0  ...   \n",
       "7945             1.0         0.008598       0.0                 1.0  ...   \n",
       "3508             1.0         0.005304       0.0                 4.0  ...   \n",
       "...              ...              ...       ...                 ...  ...   \n",
       "6564             1.0         0.003361       1.0                 1.0  ...   \n",
       "1685             0.0         0.009837       1.0                 4.0  ...   \n",
       "916              1.0         0.004938       1.0                 1.0  ...   \n",
       "43               0.0         0.016838       1.0                 2.0  ...   \n",
       "1003             0.0         0.005815       1.0                 3.0  ...   \n",
       "\n",
       "      Male  France  Germany  Spain  zero_balance  female_male_ratio  \\\n",
       "9254   1.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "1561   1.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "8829   0.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "7945   0.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "3508   0.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "...    ...     ...      ...    ...           ...                ...   \n",
       "6564   0.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "1685   0.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "916    1.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "43     0.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "1003   0.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "\n",
       "      tenure_balance_ratio  salary_balance_ratio  high_balance_low_salary  \\\n",
       "9254              0.000000              0.000000                      1.0   \n",
       "1561              0.570320              1.400882                      0.0   \n",
       "8829              0.000000              0.000000                      1.0   \n",
       "7945              0.000000              0.000000                      1.0   \n",
       "3508              0.000000              0.000000                      1.0   \n",
       "...                    ...                   ...                      ...   \n",
       "6564              0.000000              0.262062                      0.0   \n",
       "1685              0.000000              0.827415                      0.0   \n",
       "916               0.000000              0.000000                      1.0   \n",
       "43                0.259616              1.264897                      0.0   \n",
       "1003              0.292123              0.327687                      0.0   \n",
       "\n",
       "      high_salary_low_balance  \n",
       "9254                      0.0  \n",
       "1561                      1.0  \n",
       "8829                      0.0  \n",
       "7945                      0.0  \n",
       "3508                      0.0  \n",
       "...                       ...  \n",
       "6564                      0.0  \n",
       "1685                      0.0  \n",
       "916                       0.0  \n",
       "43                        1.0  \n",
       "1003                      0.0  \n",
       "\n",
       "[9532 rows x 23 columns]"
      ]
     },
     "execution_count": 508,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a408cb4",
   "metadata": {},
   "source": [
    "## Training Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 516,
   "id": "6882ed6f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 16 candidates, totalling 80 fits\n"
     ]
    }
   ],
   "source": [
    "parameters = {'model__n_estimators':[100,150], \n",
    "              'model__learning_rate':[0.1,0.2],\n",
    "              'model__max_features':[5,10],\n",
    "              'model__subsample':[1,0.9]}\n",
    "gb_model = build_model(GradientBoostingClassifier(random_state=42), parameters, excludeComplain=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 517,
   "id": "8f5ca6a2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are: {'model__learning_rate': 0.1, 'model__max_features': 5, 'model__n_estimators': 100, 'model__subsample': 1}\n",
      "The best score on training set is: 0.8268988669744021\n",
      "Accuracy on test set:  0.8585\n",
      "ROC score:  0.8697690289461975\n",
      "The classification scores are:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.92      0.91      0.91      1618\n",
      "         1.0       0.63      0.64      0.63       382\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.77      0.78      0.77      2000\n",
      "weighted avg       0.86      0.86      0.86      2000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate(gb_model, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "746c6e56",
   "metadata": {},
   "source": [
    "Comparison to without resampling model:"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "d40b3c29",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5183918",
   "metadata": {},
   "source": [
    "The ROC score has increased slighly. However, we see a slight drop in accuracy in test set. The major improvement seen is the f1-score on true class from 0.56 -> 0.63 (~11% increase)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e9ded1e",
   "metadata": {},
   "source": [
    "### Using SMOTE technique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c9b22571",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Re-initialise X and y train \n",
    "X_train = train[features].astype('float')\n",
    "y_train = train[target].astype('float')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "36ae850c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0    6355\n",
      "1.0    3177\n",
      "Name: Exited, dtype: int64\n"
     ]
    },
    {
     "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>CreditScore</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>HasCrCard</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>Complain</th>\n",
       "      <th>Satisfaction Score</th>\n",
       "      <th>...</th>\n",
       "      <th>Male</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Spain</th>\n",
       "      <th>zero_balance</th>\n",
       "      <th>female_male_ratio</th>\n",
       "      <th>tenure_balance_ratio</th>\n",
       "      <th>salary_balance_ratio</th>\n",
       "      <th>high_balance_low_salary</th>\n",
       "      <th>high_salary_low_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.010431</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.010368</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.015515</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.009610</td>\n",
       "      <td>42.0</td>\n",
       "      <td>0.006912</td>\n",
       "      <td>0.012119</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.016977</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.570320</td>\n",
       "      <td>1.400882</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.008500</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0.005184</td>\n",
       "      <td>0.011624</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.007441</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.445949</td>\n",
       "      <td>0.640097</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.008530</td>\n",
       "      <td>27.0</td>\n",
       "      <td>0.015551</td>\n",
       "      <td>0.013741</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.013261</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>1.131731</td>\n",
       "      <td>0.965060</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.007861</td>\n",
       "      <td>56.0</td>\n",
       "      <td>0.015551</td>\n",
       "      <td>0.014401</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.003421</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>1.079898</td>\n",
       "      <td>0.237542</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9527</th>\n",
       "      <td>0.010360</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.004039</td>\n",
       "      <td>0.009829</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.008981</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.077894</td>\n",
       "      <td>0.428466</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9528</th>\n",
       "      <td>0.010324</td>\n",
       "      <td>51.0</td>\n",
       "      <td>0.007355</td>\n",
       "      <td>0.010819</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.014652</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.685400</td>\n",
       "      <td>1.351617</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9529</th>\n",
       "      <td>0.009199</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0.003297</td>\n",
       "      <td>0.011715</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.012154</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.245779</td>\n",
       "      <td>1.077088</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9530</th>\n",
       "      <td>0.011276</td>\n",
       "      <td>49.0</td>\n",
       "      <td>0.013993</td>\n",
       "      <td>0.005109</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.005498</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.244677</td>\n",
       "      <td>0.077737</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9531</th>\n",
       "      <td>0.008915</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0.004871</td>\n",
       "      <td>0.007028</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.002329</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832509</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.046079</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9532 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      CreditScore   Age    Tenure   Balance  NumOfProducts  HasCrCard  \\\n",
       "0        0.010431  32.0  0.010368  0.000000            2.0        1.0   \n",
       "1        0.009610  42.0  0.006912  0.012119            2.0        1.0   \n",
       "2        0.008500  24.0  0.005184  0.011624            1.0        1.0   \n",
       "3        0.008530  27.0  0.015551  0.013741            1.0        1.0   \n",
       "4        0.007861  56.0  0.015551  0.014401            1.0        0.0   \n",
       "...           ...   ...       ...       ...            ...        ...   \n",
       "9527     0.010360  51.0  0.004039  0.009829            1.0        1.0   \n",
       "9528     0.010324  51.0  0.007355  0.010819            1.0        0.0   \n",
       "9529     0.009199  40.0  0.003297  0.011715            1.0        0.0   \n",
       "9530     0.011276  49.0  0.013993  0.005109            1.0        1.0   \n",
       "9531     0.008915  40.0  0.004871  0.007028            1.0        0.0   \n",
       "\n",
       "      IsActiveMember  EstimatedSalary  Complain  Satisfaction Score  ...  \\\n",
       "0                1.0         0.015515       0.0                 2.0  ...   \n",
       "1                1.0         0.016977       0.0                 4.0  ...   \n",
       "2                0.0         0.007441       1.0                 4.0  ...   \n",
       "3                0.0         0.013261       1.0                 2.0  ...   \n",
       "4                0.0         0.003421       1.0                 3.0  ...   \n",
       "...              ...              ...       ...                 ...  ...   \n",
       "9527             0.0         0.008981       1.0                 4.0  ...   \n",
       "9528             0.0         0.014652       1.0                 3.0  ...   \n",
       "9529             0.0         0.012154       1.0                 2.0  ...   \n",
       "9530             0.0         0.005498       1.0                 2.0  ...   \n",
       "9531             1.0         0.002329       1.0                 3.0  ...   \n",
       "\n",
       "      Male  France  Germany  Spain  zero_balance  female_male_ratio  \\\n",
       "0      1.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "1      1.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "2      1.0     0.0      0.0    1.0           0.0           0.832509   \n",
       "3      0.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "4      1.0     1.0      0.0    0.0           0.0           0.832509   \n",
       "...    ...     ...      ...    ...           ...                ...   \n",
       "9527   0.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "9528   1.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "9529   0.0     0.0      0.0    0.0           0.0           0.832509   \n",
       "9530   0.0     0.0      1.0    0.0           0.0           0.832509   \n",
       "9531   0.0     1.0      0.0    0.0           1.0           0.832509   \n",
       "\n",
       "      tenure_balance_ratio  salary_balance_ratio  high_balance_low_salary  \\\n",
       "0                 0.000000              0.000000                      1.0   \n",
       "1                 0.570320              1.400882                      0.0   \n",
       "2                 0.445949              0.640097                      0.0   \n",
       "3                 1.131731              0.965060                      0.0   \n",
       "4                 1.079898              0.237542                      0.0   \n",
       "...                    ...                   ...                      ...   \n",
       "9527              0.077894              0.428466                      1.0   \n",
       "9528              0.685400              1.351617                      0.0   \n",
       "9529              0.245779              1.077088                      0.0   \n",
       "9530              0.244677              0.077737                      0.0   \n",
       "9531              0.000000              0.046079                      1.0   \n",
       "\n",
       "      high_salary_low_balance  \n",
       "0                         0.0  \n",
       "1                         1.0  \n",
       "2                         0.0  \n",
       "3                         0.0  \n",
       "4                         0.0  \n",
       "...                       ...  \n",
       "9527                      0.0  \n",
       "9528                      0.0  \n",
       "9529                      0.0  \n",
       "9530                      0.0  \n",
       "9531                      0.0  \n",
       "\n",
       "[9532 rows x 23 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from imblearn.over_sampling import SMOTENC, SMOTE\n",
    "cat_inds = [X_train.columns.get_loc(c) for c in cat_cols]\n",
    "smote = SMOTENC(sampling_strategy=.5, categorical_features=cat_inds)\n",
    "X_resampled, y_resampled = smote.fit_resample(X_train, y_train)\n",
    "print(y_resampled.value_counts())\n",
    "X_resampled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "bc148ae7",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = X_resampled\n",
    "y_train = y_resampled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "18b10e33",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 16 candidates, totalling 80 fits\n"
     ]
    }
   ],
   "source": [
    "parameters = {'model__n_estimators':[100,150], \n",
    "              'model__learning_rate':[0.1,0.2],\n",
    "              'model__max_features':[5,10],\n",
    "              'model__subsample':[1,0.9]}\n",
    "gb_model = build_model(GradientBoostingClassifier(random_state=42), parameters, excludeComplain=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "8f912e36",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are: {'model__learning_rate': 0.1, 'model__max_features': 5, 'model__n_estimators': 100, 'model__subsample': 1}\n",
      "The best score on training set is: 0.8522870331514897\n",
      "Accuracy on test set:  0.8295\n",
      "ROC score:  0.8503881713432487\n",
      "The classification scores are:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.87      0.91      0.89      1538\n",
      "         1.0       0.65      0.56      0.60       462\n",
      "\n",
      "    accuracy                           0.83      2000\n",
      "   macro avg       0.76      0.73      0.75      2000\n",
      "weighted avg       0.82      0.83      0.82      2000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate(gb_model, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de1c25fc",
   "metadata": {},
   "source": [
    "Compared to using resampling technique, when using SMOTE the model evaluation metrics are slighlty lower.\n",
    "\n",
    "Overall, when using resampling technique, model performance slightly improves. However the improvement is not significant, so it might be not necessary to resample the data to cover the imbalanceness. For further research, I would suggest getting more data for analysis."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e7659ad0",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}