Bank_Churn_Prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Artificial Neural Network Project: Predicting Customer Churn"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "kf37d5GRU4PU"
   },
   "source": [
    "## Context\n",
    "\n",
    "An important component of sustainable success for a business is customer turnover or 'churn'. If a business better understood which aspects of their service most influence this decision (i.e., leave or stay), then management can then concentrate limited resources on mitigating turnover and retaining profitable customers."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "783YLK7Fe7yV"
   },
   "source": [
    "## Problem statement\n",
    "\n",
    "Given a database of a bank's customers, build a neural network-based classifier that can determine whether they will leave or not in the next 6 months."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IwAnuKeBgyhF"
   },
   "source": [
    "## Data description\n",
    "\n",
    "This open-source dataset contains 10,000 sample points with 14 distinct features such as CustomerId, CreditScore, Geography, Gender, Age, Tenure, Balance, etc.\n",
    "\n",
    "## Data dictionary\n",
    "\n",
    "Independent variables:\n",
    "- CustomerId: Unique ID which is assigned to each customer\n",
    "- Surname: Last name of the customer \n",
    "- CreditScore: Defines the credit history of the customer\n",
    "- Geography: A customer’s location\n",
    "- Gender: Gender of the customer   \n",
    "- Age: Age of the customer    \n",
    "- Tenure: The number of years that the customer has been with the bank\n",
    "- NumOfProducts: The number of products that a customer has purchased through the bank\n",
    "- Balance: Account balance\n",
    "- HasCrCard: A categorical variable that identifies whether the customer has a credit card or not\n",
    "- EstimatedSalary: Estimated salary \n",
    "- isActiveMember: A categorical variable that identifies whether the customer is an active member of the bank or not\n",
    "  - An active member is determined subjectively by whether a customer uses bank products regularly, makes regular transactions, etc\n",
    "\n",
    "Dependent variable:\n",
    "- Exited: A categorical variable identifying whether the customer left the bank within six months or not. It can take two values:\n",
    "  - 0=No ( Customer did not leave the bank )\n",
    "  - 1=Yes ( Customer left the bank )\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "fBYVOvFIkfdm"
   },
   "source": [
    "## Import the dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ycFapGk6GTSi"
   },
   "source": [
    "Import the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 113,
     "resources": {
      "http://localhost:8080/nbextensions/google.colab/files.js": {
       "data": "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",
       "headers": [
        [
         "content-type",
         "application/javascript"
        ]
       ],
       "ok": true,
       "status": 200,
       "status_text": ""
      }
     }
    },
    "id": "XORz4ZCYiK9e",
    "outputId": "f59303fb-3ee9-4287-cc56-ad1d11a77fb6"
   },
   "outputs": [],
   "source": [
    "#upload CSV file\n",
    "from google.colab import files\n",
    "files.upload()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "A9xMzxgjFPKC"
   },
   "source": [
    "Import libraries and settings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "hKt0pILMz_sn"
   },
   "outputs": [],
   "source": [
    "#import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import sklearn     \n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import preprocessing\n",
    "from sklearn.metrics import (\n",
    "    f1_score,\n",
    "    accuracy_score,\n",
    "    recall_score,\n",
    "    precision_score,\n",
    "    confusion_matrix,\n",
    "    roc_auc_score,\n",
    "    plot_confusion_matrix,\n",
    "    precision_recall_curve,\n",
    "    roc_curve )\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "from keras import losses\n",
    "from keras import Sequential\n",
    "from keras import backend\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.layers import BatchNormalization\n",
    "from tensorflow.keras.layers import Dropout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "id": "IB8OJgzGCEN7"
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "#format numeric data for easier readability\n",
    "pd.set_option(\n",
    "    \"display.float_format\", lambda x: \"%.2f\" % x\n",
    ")  # to display numbers rounded to 2 decimal places"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "LnGv4kkfijxW",
    "outputId": "80fa5087-5bb9-4a35-f724-9f147ee69115"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-9187428d-417d-4b23-b770-d3a0b3a2bb14\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>RowNumber</th>\n",
       "      <th>CustomerId</th>\n",
       "      <th>Surname</th>\n",
       "      <th>CreditScore</th>\n",
       "      <th>Geography</th>\n",
       "      <th>Gender</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>578</th>\n",
       "      <td>579</td>\n",
       "      <td>15734674</td>\n",
       "      <td>Lin</td>\n",
       "      <td>593</td>\n",
       "      <td>France</td>\n",
       "      <td>Female</td>\n",
       "      <td>41</td>\n",
       "      <td>6</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>65170.66</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6552</th>\n",
       "      <td>6553</td>\n",
       "      <td>15636407</td>\n",
       "      <td>Beatham</td>\n",
       "      <td>793</td>\n",
       "      <td>Germany</td>\n",
       "      <td>Female</td>\n",
       "      <td>34</td>\n",
       "      <td>5</td>\n",
       "      <td>127758.09</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>143357.03</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8718</th>\n",
       "      <td>8719</td>\n",
       "      <td>15627298</td>\n",
       "      <td>Vinogradova</td>\n",
       "      <td>589</td>\n",
       "      <td>France</td>\n",
       "      <td>Male</td>\n",
       "      <td>37</td>\n",
       "      <td>7</td>\n",
       "      <td>85146.48</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>86490.09</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7073</th>\n",
       "      <td>7074</td>\n",
       "      <td>15585855</td>\n",
       "      <td>Gould</td>\n",
       "      <td>679</td>\n",
       "      <td>France</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>16897.19</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1308</th>\n",
       "      <td>1309</td>\n",
       "      <td>15662294</td>\n",
       "      <td>Bennett</td>\n",
       "      <td>710</td>\n",
       "      <td>France</td>\n",
       "      <td>Male</td>\n",
       "      <td>33</td>\n",
       "      <td>10</td>\n",
       "      <td>118327.17</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>192928.82</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9187428d-417d-4b23-b770-d3a0b3a2bb14')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-9187428d-417d-4b23-b770-d3a0b3a2bb14 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-9187428d-417d-4b23-b770-d3a0b3a2bb14');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "      RowNumber  CustomerId      Surname  CreditScore Geography  Gender  Age  \\\n",
       "578         579    15734674          Lin          593    France  Female   41   \n",
       "6552       6553    15636407      Beatham          793   Germany  Female   34   \n",
       "8718       8719    15627298  Vinogradova          589    France    Male   37   \n",
       "7073       7074    15585855        Gould          679    France    Male   40   \n",
       "1308       1309    15662294      Bennett          710    France    Male   33   \n",
       "\n",
       "      Tenure   Balance  NumOfProducts  HasCrCard  IsActiveMember  \\\n",
       "578        6      0.00              1          1               0   \n",
       "6552       5 127758.09              1          1               0   \n",
       "8718       7  85146.48              2          1               0   \n",
       "7073       1      0.00              1          1               1   \n",
       "1308      10 118327.17              2          1               1   \n",
       "\n",
       "      EstimatedSalary  Exited  \n",
       "578          65170.66       0  \n",
       "6552        143357.03       0  \n",
       "8718         86490.09       1  \n",
       "7073         16897.19       0  \n",
       "1308        192928.82       0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#import dataset\n",
    "data = pd.read_csv('Churn.csv')\n",
    "#copy dataset to a new variable\n",
    "df = data.copy()\n",
    "#sample five random rows from dataset\n",
    "df.sample(n=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ry8J8nA4ocet",
    "outputId": "a01565e5-09fd-469e-b330-c9c11ae8b230"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.8.0\n"
     ]
    }
   ],
   "source": [
    "#verify the version of TensorFlow\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "E0ZMDNiuxnCL"
   },
   "source": [
    "## User-defined functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "id": "3Ge_SGASY96N"
   },
   "outputs": [],
   "source": [
    "# function to plot a boxplot and a histogram along the same scale\n",
    "\n",
    "def histogram_boxplot(data, feature, figsize=(16, 6), kde=False, bins=None):\n",
    "    \"\"\"\n",
    "    Combines boxplot and histogram\n",
    "\n",
    "    data: dataframe\n",
    "    feature: dataframe column\n",
    "    figsize: size of figure (default (16,6))\n",
    "    kde: whether to show the density curve (default False)\n",
    "    bins: number of bins for histogram (default None)\n",
    "    \"\"\"\n",
    "    f2, (ax_box2, ax_hist2) = plt.subplots(\n",
    "        nrows=2,  # Number of rows of the subplot grid= 2\n",
    "        sharex=True,  # x-axis will be shared among all subplots\n",
    "        gridspec_kw={\"height_ratios\": (0.25, 0.75)},\n",
    "        figsize=figsize,\n",
    "    )  # creating the 2 subplots\n",
    "    sns.boxplot(\n",
    "        data=data, x=feature, ax=ax_box2, showmeans=True, color=\"aqua\"\n",
    "    )  # boxplot will be created and a star will indicate the mean value of the column\n",
    "    sns.histplot(\n",
    "        data=data, x=feature, kde=kde, ax=ax_hist2, bins=bins, palette=\"winter\"\n",
    "    ) if bins else sns.histplot(\n",
    "        data=data, x=feature, kde=kde, ax=ax_hist2\n",
    "    )  # For histogram\n",
    "    ax_hist2.axvline(\n",
    "        data[feature].mean(), color=\"green\", linestyle=\"--\"\n",
    "    )  # Add mean to the histogram\n",
    "    ax_hist2.axvline(\n",
    "        data[feature].median(), color=\"black\", linestyle=\"-\"\n",
    "    )  # Add median to the histogram"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "Sa13Uz73cxXO"
   },
   "outputs": [],
   "source": [
    "# function to create labeled barplots\n",
    "\n",
    "def labeled_barplot(data, feature, perc=False, n=None):\n",
    "    \"\"\"\n",
    "    Barplot with percentage at the top\n",
    "\n",
    "    data: dataframe\n",
    "    feature: dataframe column\n",
    "    perc: whether to display percentages instead of count (default is False)\n",
    "    n: displays the top n category levels (default is None, i.e., display all levels)\n",
    "    \"\"\"\n",
    "\n",
    "    total = len(data[feature])  # length of the column\n",
    "    count = data[feature].nunique()\n",
    "    if n is None:\n",
    "        plt.figure(figsize=(count + 1, 5))\n",
    "    else:\n",
    "        plt.figure(figsize=(n + 1, 5))\n",
    "        \n",
    "    ax = sns.countplot(\n",
    "        data=data,\n",
    "        x=feature,\n",
    "        palette='viridis',\n",
    "        order=data[feature].value_counts().index[:n].sort_values(),\n",
    "    )\n",
    "\n",
    "    for p in ax.patches:\n",
    "        if perc == True:\n",
    "            label = \"{:.1f}%\".format(\n",
    "                100 * p.get_height() / total\n",
    "            )  # percentage of each class of the category\n",
    "        else:\n",
    "            label = p.get_height()  # count of each level of the category\n",
    "\n",
    "        x = p.get_x() + p.get_width() / 2  # width of the plot\n",
    "        y = p.get_height()  # height of the plot\n",
    "\n",
    "        ax.annotate(\n",
    "            label,\n",
    "            (x, y),\n",
    "            ha=\"center\",\n",
    "            va=\"center\",\n",
    "            size=12,\n",
    "            xytext=(0, 5),\n",
    "            textcoords=\"offset points\"\n",
    "        )  # annotate the percentage\n",
    "\n",
    "    plt.show()  # show the plot"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "1AbF7rq-N-ZD"
   },
   "source": [
    "## Overview of the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "J90FZp8-Rxnb",
    "outputId": "21508af6-cc85-494f-a743-a5538a50a0b5"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 14)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#observe the shape of the dataframe\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "gtqyQNltMlHX",
    "outputId": "75fd0ae9-c83e-4e85-dbd8-eddfd1b77a8c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 10000 entries, 0 to 9999\n",
      "Data columns (total 14 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   RowNumber        10000 non-null  int64  \n",
      " 1   CustomerId       10000 non-null  int64  \n",
      " 2   Surname          10000 non-null  object \n",
      " 3   CreditScore      10000 non-null  int64  \n",
      " 4   Geography        10000 non-null  object \n",
      " 5   Gender           10000 non-null  object \n",
      " 6   Age              10000 non-null  int64  \n",
      " 7   Tenure           10000 non-null  int64  \n",
      " 8   Balance          10000 non-null  float64\n",
      " 9   NumOfProducts    10000 non-null  int64  \n",
      " 10  HasCrCard        10000 non-null  int64  \n",
      " 11  IsActiveMember   10000 non-null  int64  \n",
      " 12  EstimatedSalary  10000 non-null  float64\n",
      " 13  Exited           10000 non-null  int64  \n",
      "dtypes: float64(2), int64(9), object(3)\n",
      "memory usage: 1.1+ MB\n"
     ]
    }
   ],
   "source": [
    "#observe the data types, non-null values, and column names of the dataframe\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "mjbzixE9ONdD"
   },
   "outputs": [],
   "source": [
    "#drop column RowNumber, CustomerId, and Surname from the dataset\n",
    "df.drop(['RowNumber','CustomerId', 'Surname'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "-3IfMkkvzE9u"
   },
   "source": [
    "- *RowNumber*, *Customer_ID*, and *Surname* values should not provide any material information and will not be included in training the network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Aw_rnJ7RTA3b",
    "outputId": "a0424c5d-b365-423d-ca99-77bfcb4f529b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CreditScore        0\n",
       "Geography          0\n",
       "Gender             0\n",
       "Age                0\n",
       "Tenure             0\n",
       "Balance            0\n",
       "NumOfProducts      0\n",
       "HasCrCard          0\n",
       "IsActiveMember     0\n",
       "EstimatedSalary    0\n",
       "Exited             0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#check for duplicate values\n",
    "df[df.duplicated()].count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "id": "jPecOUiGQsMH"
   },
   "outputs": [],
   "source": [
    "#convert all columns with dtype object into category\n",
    "for col in df.columns[df.dtypes=='object']:\n",
    "    df[col] = df[col].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "uebbsEAsGi_c"
   },
   "outputs": [],
   "source": [
    "#convert all columns with dtype int64 into int32\n",
    "for col in df.columns[df.dtypes=='int64']:\n",
    "    df[col] = df[col].astype('int32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "id": "A_8WX5COGjkM"
   },
   "outputs": [],
   "source": [
    "#convert all columns with dtype float64 into float32\n",
    "for col in df.columns[df.dtypes=='float64']:\n",
    "    df[col] = df[col].astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "1pC4fDmya8W4"
   },
   "outputs": [],
   "source": [
    "int_cat = ['HasCrCard','IsActiveMember']\n",
    "#convert all columns in list to category\n",
    "for col in int_cat:\n",
    "    df[col] = df[col].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "aifvcc1ZO0Gs",
    "outputId": "5e41bd24-435f-4b21-b794-3de78e78771e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 10000 entries, 0 to 9999\n",
      "Data columns (total 11 columns):\n",
      " #   Column           Non-Null Count  Dtype   \n",
      "---  ------           --------------  -----   \n",
      " 0   CreditScore      10000 non-null  int32   \n",
      " 1   Geography        10000 non-null  category\n",
      " 2   Gender           10000 non-null  category\n",
      " 3   Age              10000 non-null  int32   \n",
      " 4   Tenure           10000 non-null  int32   \n",
      " 5   Balance          10000 non-null  float32 \n",
      " 6   NumOfProducts    10000 non-null  int32   \n",
      " 7   HasCrCard        10000 non-null  category\n",
      " 8   IsActiveMember   10000 non-null  category\n",
      " 9   EstimatedSalary  10000 non-null  float32 \n",
      " 10  Exited           10000 non-null  int32   \n",
      "dtypes: category(4), float32(2), int32(5)\n",
      "memory usage: 313.1 KB\n"
     ]
    }
   ],
   "source": [
    "#verify conversions\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "fdWcJYFDT-dx"
   },
   "source": [
    "#### Observations\n",
    "- The dataframe has 10000 rows and 11 columns.\n",
    "- There are no duplicate values in the dataset.\n",
    "- There are no null values across all columns.\n",
    "- There are 4 columns of the dtype category, 2 columns of the dtype float64, and 5 columns of the dtype int64.\n",
    "- The target variable for the Neural Network will be *Exited*."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "b1CNQabJX8XL"
   },
   "source": [
    "## Univariate analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "2Fh2gfNkZEbj"
   },
   "source": [
    "### Summary of float variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 758
    },
    "id": "2lCrAugIZDdJ",
    "outputId": "f8f6cc60-df58-4d39-9930-8b47a60ba6be"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7kAAAFzCAYAAAD7S847AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de7BlVX0v+u9PXg1BBaRD2qaxwRAS46nTmvZFLEPUIFKx0BwjpIyC8QAn6i2pa6yAScrE90lpfNyrBDhywccVOD7KhqAIKLFSKNKYlqctbSN2tyAvxQcPeYz7x57N3bS9m92w1p57z/X5VM3ac4055ly/1Yxem2+Pucaq1loAAABgCB7XdwEAAAAwKkIuAAAAgyHkAgAAMBhCLgAAAIMh5AIAADAYQi4AAACDsWPfBYzD3nvv3ZYvX953GQAAAIzBFVdccVtrbfHWjg0y5C5fvjyrV6/uuwwAAADGoKpunOmY25UBAAAYDCEXAACAwRByAQAAGAwhFwAAgMEY5MJT891pp52W9evX910GwNjcdNNNSZIlS5b0XAlJcsABB+TYY4/tuwwAmBNCbg/Wr1+f73z/+7ln6dK+SwEYi0W//GWS5If33NNzJSzatKnvEgBgTgm5Pbln6dLc8OY3910GwFjs/+EPJ4n3uXlg838LAJgUPpMLAADAYAi5AAAADIaQCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMhpALAADAYAi5AAAADIaQCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMxo59FzCJbrrppux8//19lwEAAPBrTjvttCTJscce23Mlj46Q24O77747j3vwwb7LAAAA+DXr16/vu4THxO3KAAAADIaQCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMhpALAADAYAi5AAAADIaQCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMhpALMCJ17x1ZtObE1K9+0ncpAAATS8gFGJGdbjwrO9x5bXa68ay+SwEAmFg79l0AwBDUvXdkpx9fnErLTjdflPueclTaznv2XRZk51tvzfr77stJJ53UdykALBDr16/Prrvu2ncZj9pgZnKr6riqWl1Vq2+99da+ywEmzE43npW0B6cetAfN5gIA9GQwM7mttVOTnJokK1eubD2XA0yQh2Zx2/1Tj9v9ZnOZN361eHEOWLQo733ve/suBYAFYqHf/TOYmVyAvjxsFnczs7kAAL0QcgEeox1+9t2HZnE3q3Z/drjzup4qAgCYXIO5XRmgL3ev/EjfJQAA0DGTCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMhpALAADAYAi5AAAADIaQCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMxo59FzCJdt1119x5//19lwEAAPBrDjjggL5LeEyE3B4sWbIkP7znnr7LAAAA+DXHHnts3yU8Jm5XBgAAYDCEXAAAAAZDyAUAAGAwhFwAAAAGQ8gFAABgMIRcAAAABkPIBQAAYDCEXAAAAAZDyAUAAGAwhFwAAAAGQ8gFAABgMHbsu4BJtWjTpuz/4Q/3XQbAWCzauDFJvM/NA4s2bUqe+tS+ywCAOSPk9uCAAw7ouwSAsbrpN34jSbJk0aKeKyFPfarfOwBMFCG3B8cee2zfJQAAAAySz+QCAAAwGEIuAAAAgyHkAgAAMBhCLgAAAIMh5AIAADAYQi4AAACDUa21vmsYuaq6NcmNfdfxCPZOclvfRTDxjEPmA+OQvhmDzAfGIfPBQhqHT2mtLd7agUGG3IWgqla31lb2XQeTzThkPjAO6ZsxyHxgHDIfDGUcul0ZAACAwRByAQAAGAwhtz+n9l0AxDhkfpjYcVhVD1TVmqr6TlV9u6oOnsU5v5iL2ibMxI5B5hXjkPlgEOPQZ3IBoCdV9YvW2u7d/kuSvK219kezPQcA+HVmcgFgfnhCkp8kSVXtXlUXd7O7V1XVEVt2nqlPVS2vquuq6rSquqaqvlJVu3bHfruqLpo2c/zUrv2tVXV5VV1ZVf80h68ZAEbOTC4A9KSqHkhyVZJFSZYkeWFr7Yqq2jHJbq21n1XV3km+meTA1lrbPJM7U58kT0myLsnK1tqaqjonyarW2qeq6rIk72utfaGqFmXqH7ufn+SVSY5PUklWJfnn1trX5/LPAgBGZce+CwCACXZ3a21FklTV85J8oqqenqmw+Z6qekGSB5MsTbJPkpunnTtTnyS5obW2ptu/Isnyqnp8kqWttS8kSWvtnu55D01yaJL/7PrvnqmwLOQCsCAJuQAwD7TWvtHNyC5Ocnj38w9aa/dV1Q8yNds73au30efeaf0eSLLrNp66kry3tXbKY38VANA/n8kFgHmgqn43yQ5Jbk/yxCS3dOH1jzN1C/KWZtPnIa21nyfZWFUv755vl6raLckFSf6qqjYvgLW0qn5zZC8MAOaYmVwA6M+uVbX5tuJKcnRr7YGq+nSSc6vqqiSrk3x3K+fOps+WXpPklKp6R5L7kvx5a+0rVfV7Sb5RVUnyiyR/meSWx/LCAKAvFp4CAABgMNyuDAAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMhpALAADAYAi5AAAADIaQCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMhpALAADAYAi5AAAADIaQCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMxo59FzAOe++9d1u+fHnfZQAT5s577kySPHHRE3uupH9r165Nkhx00EE9VwIADNEVV1xxW2tt8daODTLkLl++PKtXr+67DICJdcghhyRJLrnkkl7rAACGqapunOmY25UBRmTtbWuz9ra1fZcBADDRBjmTC9CH4887PklyyTGX9FsIAMAEM5MLAADAYAi5AAAADIaQCwAAwGAIuQAAAAyGhacARuTvX/D3fZcAADDxhFyAEXnxAS/uuwQAgInndmWAEVlz85qsuXlN32UAAEw0M7k9Wbpsv/xo44a+y3iYJ++7LJs2/LDvMmDBOuHLJyTxPbkAAH0Scnvyo40bcuQpl/ZdxsOcffzBfZcAAADwmLhdGQAAgMEYW8itqkVV9a2q+k5VXVNV/9S1n1FVN1TVmm5b0bVXVX2kqtZV1ZVV9cxp1zq6qq7vtqPHVTMAAAAL2zhvV743yQtba7+oqp2S/EdVfak79tbW2me36P/SJAd223OSnJzkOVW1V5K3J1mZpCW5oqpWtdZ+MsbaAQAAWIDGFnJbay3JL7qHO3Vb28YpRyT5RHfeN6tqj6pakuSQJBe21u5Ikqq6MMlhST4zrtoBHo33vOg9fZcAADDxxvqZ3KraoarWJLklU0H1su7Qu7tbkj9YVbt0bUuTTF9ueGPXNlM7wLxy8LKDc/AyC7gBAPRprCG3tfZAa21Fkn2TPLuqnp7kpCS/m+RZSfZK8rejeK6qOq6qVlfV6ltvvXUUlwTYLpduuDSXbphfq6YDAEyaOVldubX20yRfS3JYa+2mNuXeJP9Pkmd33TYlWTbttH27tpnat3yOU1trK1trKxcvXjyOlwGwTW+7+G1528Vv67sMAICJNs7VlRdX1R7d/q5J/iTJd7vP2aaqKsnLk1zdnbIqyWu7VZafm+TO1tpNSS5IcmhV7VlVeyY5tGsDAACAhxnn6spLkpxZVTtkKkyf01o7r6q+WlWLk1SSNUn+R9f//CSHJ1mX5K4kr0uS1todVfXOJJd3/d6xeREqAAAAmG6cqytfmeQZW2l/4Qz9W5I3znDs9CSnj7RAAAAABmdOPpMLAAAAc2GctysDTJQPHfahvksAAJh4Qi7AiKz4rRV9lwAAMPHcrgwwIhetvygXrb+o7zIAACaamVyAEXnX19+VJHnxAS/uuRIAgMllJhcAAIDBEHIBAAAYDCEXAACAwRByAQAAGAwLTwGMyCl/ekrfJQAATDwhF2BEDtr7oL5LAACYeG5XBhiRc9eem3PXntt3GQAAE81MLsCIfOAbH0iSvOygl/VcCQDA5DKTCwAAwGAIuQAAAAyGkAsAAMBgCLkAAAAMhoWnAEbkk6/4ZN8lAABMPCEXYESWPXFZ3yUAAEw8tysDjMjZV5+ds68+u+8yAAAmmplcgBE5efXJSZIjn35kz5UAAEyusc3kVtWiqvpWVX2nqq6pqn/q2vevqsuqal1VnV1VO3ftu3SP13XHl0+71kld+9qqesm4agYAAGBhG+ftyvcmeWFr7b8mWZHksKp6bpL/meSDrbXfTvKTJK/v+r8+yU+69g92/VJVT0tyVJLfT3JYko9V1Q5jrBsAAIAFamwht035Rfdwp25rSV6Y5LNd+5lJXt7tH9E9Tnf8RVVVXftZrbV7W2s3JFmX5NnjqhsAAICFa6wLT1XVDlW1JsktSS5M8v0kP22t3d912Zhkabe/NMmGJOmO35nkSdPbt3LO9Oc6rqpWV9XqW2+9dRwvBwAAgHlurAtPtdYeSLKiqvZI8oUkvzvG5zo1yalJsnLlyjau5wGYyWdf9dlH7gQAwFjNyerKrbWfVtXXkjwvyR5VtWM3W7tvkk1dt01JliXZWFU7JnliktuntW82/RyAeWPv3fbuuwQAgIk3ztWVF3czuKmqXZP8SZLrknwtySu7bkcn+WK3v6p7nO74V1trrWs/qlt9ef8kByb51rjqBni0zlhzRs5Yc0bfZQAATLRxzuQuSXJmtxLy45Kc01o7r6quTXJWVb0ryX8m+XjX/+NJPllV65LckakVldNau6aqzklybZL7k7yxuw0aYF7ZHHCPWXFMr3UAAEyysYXc1tqVSZ6xlfb12crqyK21e5L8+QzXeneSd4+6RgAAAIZlrKsrAwAAwFwScgEAABgMIRcAAIDBmJOvEAKYBOe/+vy+SwAAmHhCLsCI7LbTbn2XAAAw8dyuDDAiH7v8Y/nY5R/ruwwAgIkm5AKMyDnXnJNzrjmn7zIAACaakAsAAMBgCLkAAAAMhpALAADAYAi5AAAADIavEAIYkUuOuaTvEgAAJp6ZXAAAAAZDyAUYkfdf+v68/9L3910GAMBEE3IBRuS8752X8753Xt9lAABMNCEXAACAwRByAQAAGAwhFwAAgMHwFUIAI7LrTrv2XQIAwMQTcgFG5Euv/lLfJQAATLyx3a5cVcuq6mtVdW1VXVNVb+7a/7GqNlXVmm47fNo5J1XVuqpaW1UvmdZ+WNe2rqpOHFfNAAAALGzjnMm9P8lbWmvfrqrHJ7miqi7sjn2wtfawL5OsqqclOSrJ7yd5cpKLqup3usMfTfInSTYmubyqVrXWrh1j7QDb7Z3//s4kyT/80T/0XAkAwOQa20xua+2m1tq3u/2fJ7kuydJtnHJEkrNaa/e21m5Isi7Js7ttXWttfWvtV0nO6voCzCsX33BxLr7h4r7LAACYaHOyunJVLU/yjCSXdU1vqqorq+r0qtqza1uaZMO00zZ2bTO1b/kcx1XV6qpafeutt474FQAAALAQjD3kVtXuST6X5ITW2s+SnJzkqUlWJLkpyQdG8TyttVNbaytbaysXL148iksCAACwwIx1deWq2ilTAffTrbXPJ0lr7cfTjp+W5Lzu4aYky6advm/Xlm20AwAAwEPGubpyJfl4kutaa/8yrX3JtG6vSHJ1t78qyVFVtUtV7Z/kwCTfSnJ5kgOrav+q2jlTi1OtGlfdAI/Wk3Z7Up6025P6LgMAYKKNcyb3D5O8JslVVbWma3tbkr+oqhVJWpIfJDk+SVpr11TVOUmuzdTKzG9srT2QJFX1piQXJNkhyemttWvGWDfAo/K5V32u7xIAACbe2EJua+0/ktRWDp2/jXPeneTdW2k/f1vnAQAAQDJHqysDTIKTLjopJ110Ut9lAABMtLEuPAUwSb6x8Rt9lwAAMPHM5AIAADAYQi4AAACDIeQCAAAwGD6TCzAi+z5h375LAACYeEIuwIh86s8+1XcJAAATz+3KAAAADIaQCzAiJ3z5hJzw5RP6LgMAYKK5XRlgRNbcvKbvEgAAJp6ZXAAAAAZDyAUAAGAwhFwAAAAGw2dyAUbkd570O32XAAAw8YRcgBE59WWn9l0CAMDEc7syAAAAgyHkAozIcecel+POPa7vMgAAJtqsQm5V/eFs2gAm2fdu/16+d/v3+i4DAGCizXYm9/+aZRsAAAD0ZpsLT1XV85IcnGRxVf2f0w49IckO4ywMAAAAttcjzeTunGT3TIXhx0/bfpbklds6saqWVdXXquraqrqmqt7cte9VVRdW1fXdzz279qqqj1TVuqq6sqqeOe1aR3f9r6+qox/9ywUAAGDItjmT21r79yT/XlVntNZu3M5r35/kLa21b1fV45NcUVUXJjkmycWttfdV1YlJTkzyt0lemuTAbntOkpOTPKeq9kry9iQrk7TuOqtaaz/ZznoAxmrFb63ouwQAgIk32+/J3aWqTk2yfPo5rbUXznRCa+2mJDd1+z+vquuSLE1yRJJDum5nJrkkUyH3iCSfaK21JN+sqj2qaknX98LW2h1J0gXlw5J8Zpa1A8yJDx32ob5LAACYeLMNuf87yb8m+V9JHtjeJ6mq5UmekeSyJPt0AThJbk6yT7e/NMmGaadt7NpmagcAAICHmW3Ivb+1dvKjeYKq2j3J55Kc0Fr7WVU9dKy11qqqPZrrbuV5jktyXJLst99+o7gkwHb5y8//ZZLkU3/2qZ4rAQCYXLP9CqFzq+oNVbWkWzhqr+6zsttUVTtlKuB+urX2+a75x91tyOl+3tK1b0qybNrp+3ZtM7U/TGvt1NbaytbaysWLF8/yZQGMzsafbczGn23suwwAgIk225B7dJK3Jrk0yRXdtnpbJ9TUlO3Hk1zXWvuXaYdWddfbfN0vTmt/bbfK8nOT3Nnd1nxBkkOras9uJeZDuzYAAAB4mFndrtxa2/9RXPsPk7wmyVVVtaZre1uS9yU5p6pen+TGJK/qjp2f5PAk65LcleR13XPfUVXvTHJ51+8dmxehAgAAgOlmFXKr6rVba2+tfWKmc1pr/5GkZjj8oq30b0neOMO1Tk9y+iNXCgAAwCSb7cJTz5q2vyhTIfXbSWYMuQCT5nn7Pq/vEgAAJt5sb1f+P6Y/rqo9kpw1looAFqj3vvi9fZcAADDxZrvw1JZ+meTRfE4XAAAAxma2n8k9N8nm77PdIcnvJTlnXEUBLET/7Zz/liT53Ks+13MlAACTa7afyX3/tP37k9zYWvNlkADT3H7X7X2XAAAw8WZ1u3Jr7d+TfDfJ45PsmeRX4ywKAAAAHo1ZhdyqelWSbyX580x9r+1lVfXKcRYGAAAA22u2tyv/XZJntdZuSZKqWpzkoiSfHVdhAAAAsL1mG3Iftzngdm7Po1+ZGWCQXrT/i/ouAQBg4s025H65qi5I8pnu8ZFJzh9PSQAL0z/80T/0XQIAwMTbZsitqt9Osk9r7a1V9WdJnt8d+kaST4+7OAAAANgejzST+6EkJyVJa+3zST6fJFX1X7pjLxtrdQALyEs//dIkyZde/aWeKwEAmFyPFHL3aa1dtWVja+2qqlo+looAFqi777u77xIAACbeIy0etcc2ju06ykIAAADgsXqkkLu6qo7dsrGq/nuSK8ZTEgAAADw6j3S78glJvlBVr87/H2pXJtk5ySvGWRgAAABsr22G3Nbaj5McXFV/nOTpXfO/tda+OvbKABaYP/2dP+27BACAiTer78ltrX0tydfGXAvAgvY3B/9N3yUAAEy8R/pMLgAAACwYQi7AiBxyxiE55IxD+i4DAGCiCbkAAAAMxthCblWdXlW3VNXV09r+sao2VdWabjt82rGTqmpdVa2tqpdMaz+sa1tXVSeOq14AAAAWvnHO5J6R5LCttH+wtbai285Pkqp6WpKjkvx+d87HqmqHqtohyUeTvDTJ05L8RdcXAAAAfs2sVld+NFprX6+q5bPsfkSSs1pr9ya5oarWJXl2d2xda219klTVWV3fa0dcLgAAAAMwtpC7DW+qqtcmWZ3kLa21nyRZmuSb0/ps7NqSZMMW7c/Z2kWr6rgkxyXJfvvtN+qaAR7Rq37/VX2XAAAw8eZ64amTkzw1yYokNyX5wKgu3Fo7tbW2srW2cvHixaO6LMCsveFZb8gbnvWGvssAAJhoczqT21r78eb9qjotyXndw01Jlk3rum/Xlm20A8wrd913V5Jkt51267kSAIDJNaczuVW1ZNrDVyTZvPLyqiRHVdUuVbV/kgOTfCvJ5UkOrKr9q2rnTC1OtWouawaYrcM/fXgO//Thj9wRAICxGdtMblV9JskhSfauqo1J3p7kkKpakaQl+UGS45OktXZNVZ2TqQWl7k/yxtbaA9113pTkgiQ7JDm9tXbNuGoGAABgYRvn6sp/sZXmj2+j/7uTvHsr7ecnOX+EpQEAADBQc73wFAAAAIyNkAsAAMBg9PE9uQCDdMyKY/ouAQBg4gm5ACMi5AIA9M/tygAjcttdt+W2u27ruwwAgIlmJhdgRF55ziuTJJccc0m/hQAATDAzuQAAAAyGkAsAAMBgCLkAAAAMhpALAADAYFh4CmBE/nrlX/ddAgDAxBNyAUbkyKcf2XcJAAATz+3KACOy4c4N2XDnhr7LAACYaGZyAUbkNV94TRLfkwsA0CczuQAAAAyGkAsAAMBgCLkAAAAMhpALAADAYFh4CmBE3vK8t/RdAgDAxBNyAUbkZQe9rO8SAAAm3thuV66q06vqlqq6elrbXlV1YVVd3/3cs2uvqvpIVa2rqiur6pnTzjm66399VR09rnoBHqu1t63N2tvW9l0GAMBEG+dncs9IctgWbScmubi1dmCSi7vHSfLSJAd223FJTk6mQnGStyd5TpJnJ3n75mAMMN8cf97xOf684/suAwBgoo0t5LbWvp7kji2aj0hyZrd/ZpKXT2v/RJvyzSR7VNWSJC9JcmFr7Y7W2k+SXJhfD84AAACQZO5XV96ntXZTt39zkn26/aVJNkzrt7Frm6n911TVcVW1uqpW33rrraOtGgAAgAWht68Qaq21JG2E1zu1tbaytbZy8eLFo7osAPyapcv2S1XNq23psv36/mMBgHlhrldX/nFVLWmt3dTdjnxL174pybJp/fbt2jYlOWSL9kvmoE4AmNGPNm7Ikadc2ncZD3P28Qf3XQIAzAtzHXJXJTk6yfu6n1+c1v6mqjorU4tM3dkF4QuSvGfaYlOHJjlpjmsGmJW/f8Hf910CAMDEG1vIrarPZGoWdu+q2pipVZLfl+Scqnp9khuTvKrrfn6Sw5OsS3JXktclSWvtjqp6Z5LLu37vaK1tuZgVwLzw4gNe3HcJg7R02X750cYNj9wRACBjDLmttb+Y4dCLttK3JXnjDNc5PcnpIywNYCzW3LwmSbLit1b0XMmwuDUYANgec327MsBgnfDlE5IklxxzSb+FMJket2Oqqu8qHubJ+y7Lpg0/7LsMACaMkAsAQ/Dg/Wa8ASA9foUQAAAAjJqQCwAAwGAIuQAAAAyGz+QCjMh7XvSevksAAJh4Qi7AiBy8zCI7AAB9c7sywIhcuuHSXLphfq1uCwAwaczkAozI2y5+WxLfkwsA0CczuQAAAAyGmVwAYDwet2Oqqu8qHubJ+y7Lpg0/7LsMAMZIyAXgIUuX7ZcfbdwwsuvNt4DDHHvw/hx5yvz6nPrZx1sgDmDohFwAHvKjjRtGEkq++oE3Jkle+JaPPuZrCSUAwPYQcgFG5EOHfajvEgAAJp6QCzAiK35rRd8lAABMPKsrA4zIResvykXrL+q7DACAiWYmF2BE3vX1dyVJXnzAi3uuBABgcpnJBQAAYDCEXAAAAAZDyAUAAGAwegm5VfWDqrqqqtZU1equba+qurCqru9+7tm1V1V9pKrWVdWVVfXMPmoGAABg/utz4ak/bq3dNu3xiUkubq29r6pO7B7/bZKXJjmw256T5OTuJ8C8csqfntJ3CQAAE28+3a58RJIzu/0zk7x8Wvsn2pRvJtmjqpb0USDAthy090E5aO+D+i4DAGCi9RVyW5KvVNUVVXVc17ZPa+2mbv/mJPt0+0uTbJh27sauDWBeOXftuTl37bl9lwEAMNH6ul35+a21TVX1m0kurKrvTj/YWmtV1bbngl1YPi5J9ttvv9FVCjBLH/jGB5IkLzvoZT1XAgAwuXqZyW2tbep+3pLkC0meneTHm29D7n7e0nXflGTZtNP37dq2vOaprbWVrbWVixcvHmf5ACOxdNl+qap5tcHgPW7H3v+ebbktXeYf5wFGac5ncqvqN5I8rrX2827/0CTvSLIqydFJ3tf9/GJ3yqokb6qqszK14NSd025rBliwfrRxQ4485dK+y3iYs48/uO8SYLwevN/fO4CB6+N25X2SfKGbMdgxyf/bWvtyVV2e5Jyqen2SG5O8qut/fpLDk6xLcleS1819yQAAACwEcx5yW2vrk/zXrbTfnuRFW2lvSd44B6UBAACwwPX5PbkAg/LJV3yy7xIAACaekAswIsueuOyROwEAMFZ9fU8uwOCcffXZOfvqs/suAwBgopnJBRiRk1efnCQ58ulH9lwJAMDkMpMLAADAYAi5AAAADIaQCwAAwGD4TC4wEZYu2y8/2rhhvE9yzNSPel2N93mAYXncjqmaP+8bT953WTZt+GHfZQA8akIuMBF+tHFDjjzl0rE+x70P/jRJssvz9phV/7OPP3ic5QALxYP3j/39aXt4bwIWOiEXYER2edzswi0AAOPjM7kAI3LDXf+WG+76t77LAACYaEIuwIjccPf5ueHu8/suAwBgogm5AAAADIaQCwAAwGBYeAoYuTn5uh4AxmOefaVR4muNgO0j5AIjNxdf17O9fCUGwCzNs680SryHA9tHyAUYkRfs9YG+SwAAmHhCLsCI7FiL+i4BAGDiWXgKYESu/+Xncv0vP9d3GQAAE03IBRiRDfd8NRvu+WrfZQAMT7cY1nzali7br+8/FWAGC+Z25ao6LMmHk+yQ5H+11t7Xc0kwb1jNGIBBsxgWsB0WRMitqh2SfDTJnyTZmOTyqlrVWru238pgfphvqxn7xQ8AQF8WRMhN8uwk61pr65Okqs5KckQSIZc5Z9YUAPB9wjB/LZSQuzTJ9FSxMclzeqqFOTRfA+V8mjVNzJwCwJybj7dQ/3KJgh0AAAfTSURBVPULBG9IUq21vmt4RFX1yiSHtdb+e/f4NUme01p707Q+xyU5rnt4UJK1c17o9tk7yW19F8HEMw6ZD4xD+mYMMh8Yh8wHC2kcPqW1tnhrBxbKTO6mJMumPd63a3tIa+3UJKfOZVGPRVWtbq2t7LsOJptxyHxgHNI3Y5D5wDhkPhjKOFwoXyF0eZIDq2r/qto5yVFJVvVcEwAAAPPMgpjJba3dX1VvSnJBpr5C6PTW2jU9lwUAAMA8syBCbpK01s5Pcn7fdYzQgrm1mkEzDpkPjEP6ZgwyHxiHzAeDGIcLYuEpAAAAmI2F8plcAAAAeERCbg+q6rCqWltV66rqxL7rYeGrqh9U1VVVtaaqVndte1XVhVV1ffdzz669quoj3fi7sqqeOe06R3f9r6+qo6e1/0F3/XXdufPrS/joRVWdXlW3VNXV09rGPu5meg4m0wzj8B+ralP3nrimqg6fduykbkytraqXTGvf6u/mbtHLy7r2s7sFMFNVu3SP13XHl8/NK2a+qaplVfW1qrq2qq6pqjd37d4PmTPbGIeT+X7YWrPN4ZaphbO+n+SAJDsn+U6Sp/Vdl21hb0l+kGTvLdr+OcmJ3f6JSf5nt394ki8lqSTPTXJZ175XkvXdzz27/T27Y9/q+lZ37kv7fs22/rckL0jyzCRXT2sb+7ib6Tlsk7nNMA7/McnfbKXv07rfu7sk2b/7fbzDtn43JzknyVHd/r8m+etu/w1J/rXbPyrJ2X3/Wdj62ZIsSfLMbv/xSb7XjTXvh7b5MA4n8v3QTO7ce3aSda219a21XyU5K8kRPdfEMB2R5Mxu/8wkL5/W/ok25ZtJ9qiqJUlekuTC1todrbWfJLkwyWHdsSe01r7Zpt69PjHtWkyw1trXk9yxRfNcjLuZnoMJNMM4nMkRSc5qrd3bWrshybpM/V7e6u/mbrbshUk+252/5ZjePA4/m+RF7nKZTK21m1pr3+72f57kuiRL4/2QObSNcTiTQb8fCrlzb2mSDdMeb8y2ByDMRkvylaq6oqqO69r2aa3d1O3fnGSfbn+mMbit9o1baYetmYtxN9NzwHRv6m4FPX3aLZzbOw6flOSnrbX7t2h/2LW643d2/Zlg3W2az0hyWbwf0pMtxmEyge+HQi4Mw/Nba89M8tIkb6yqF0w/2P3Lr6XUmVNzMe6MbWZwcpKnJlmR5KYkH+i3HCZBVe2e5HNJTmit/Wz6Me+HzJWtjMOJfD8UcufepiTLpj3et2uDR621tqn7eUuSL2TqVpMfd7c4pft5S9d9pjG4rfZ9t9IOWzMX426m54AkSWvtx621B1prDyY5LVPvicn2j8PbM3Ur6Y5btD/sWt3xJ3b9mUBVtVOmgsWnW2uf75q9HzKntjYOJ/X9UMide5cnObBbnWznTH04e1XPNbGAVdVvVNXjN+8nOTTJ1ZkaV5tXZjw6yRe7/VVJXtut7vjcJHd2tzpdkOTQqtqzu5Xl0CQXdMd+VlXP7T5f8dpp14ItzcW4m+k5IMlD/7O/2Ssy9Z6YTI2do7qVQPdPcmCmFvTZ6u/mbmbsa0le2Z2/5ZjePA5fmeSrXX8mTPce9fEk17XW/mXaIe+HzJmZxuHEvh/2vfLVJG6ZWlXve5lauezv+q7HtrC3TK1+951uu2bzmMrUZyEuTnJ9kouS7NW1V5KPduPvqiQrp13rrzK18MC6JK+b1r4yU2+K30/yfyepvl+3rf8tyWcydevTfZn6bM7r52LczfQctsncZhiHn+zG2ZWZ+p+vJdP6/103ptZm2krxM/1u7t5jv9WNz/+dZJeufVH3eF13/IC+/yxsvY3B52fqNuErk6zptsO9H9rmctvGOJzI98PNf0EAAABgwXO7MgAAAIMh5AIAADAYQi4AAACDIeQCAAAwGEIuAAAAgyHkAkBPquqBqlpTVd+pqm9X1cGzOOcXc1EbACxUO/ZdAABMsLtbayuSpKpekuS9Sf6o35IAYGEzkwsA88MTkvwkSapq96q6uJvdvaqqjtiy80x9qmp5VV1XVadV1TVV9ZWq2rU79ttVddG0meOndu1vrarLq+rKqvqnOXzNADBy1VrruwYAmEhV9UCSq5IsSrIkyQtba1dU1Y5Jdmut/ayq9k7yzSQHttZaVf2itbb7TH2SPCXJuiQrW2trquqcJKtaa5+qqsuSvK+19oWqWpSpf+x+fpJXJjk+SSVZleSfW2tfn8s/CwAYFbcrA0B/pt+u/Lwkn6iqp2cqbL6nql6Q5MEkS5Psk+TmaefO1CdJbmitren2r0iyvKoen2Rpa+0LSdJau6d73kOTHJrkP7v+u2cqLAu5ACxIQi4AzAOttW90M7KLkxze/fyD1tp9VfWDTM32TvfqbfS5d1q/B5Lsuo2nriTvba2d8thfBQD0z2dyAWAeqKrfTbJDktuTPDHJLV14/eNM3YK8pdn0eUhr7edJNlbVy7vn26WqdktyQZK/qqrdu/alVfWbI3thADDHzOQCQH92rarNtxVXkqNbaw9U1aeTnFtVVyVZneS7Wzl3Nn229Jokp1TVO5Lcl+TPW2tfqarfS/KNqkqSXyT5yyS3PJYXBgB9sfAUAAAAg+F2ZQAAAAZDyAUAAGAwhFwAAAAGQ8gFAABgMIRcAAAABkPIBQAAYDCEXAAAAAZDyAUAAGAw/j9KHx4eMzQWOwAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 1152x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1152x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot all float32 variables\n",
    "for col in df.columns[df.dtypes == 'float32']:\n",
    "    histogram_boxplot(df, col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 112
    },
    "id": "mC4fU_ihYJZH",
    "outputId": "353e25d0-551d-4bb3-d7b0-36315a1b4d49"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-fa80a942-042c-4591-b4ce-801121cd4bdb\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>Balance</th>\n",
       "      <td>10000.00</td>\n",
       "      <td>76485.78</td>\n",
       "      <td>62398.95</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>97198.54</td>\n",
       "      <td>127644.24</td>\n",
       "      <td>250898.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <td>10000.00</td>\n",
       "      <td>100090.12</td>\n",
       "      <td>57510.47</td>\n",
       "      <td>11.58</td>\n",
       "      <td>51002.11</td>\n",
       "      <td>100193.91</td>\n",
       "      <td>149388.25</td>\n",
       "      <td>199992.48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-fa80a942-042c-4591-b4ce-801121cd4bdb')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-fa80a942-042c-4591-b4ce-801121cd4bdb button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-fa80a942-042c-4591-b4ce-801121cd4bdb');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "                   count      mean      std   min      25%       50%  \\\n",
       "Balance         10000.00  76485.78 62398.95  0.00     0.00  97198.54   \n",
       "EstimatedSalary 10000.00 100090.12 57510.47 11.58 51002.11 100193.91   \n",
       "\n",
       "                      75%       max  \n",
       "Balance         127644.24 250898.09  \n",
       "EstimatedSalary 149388.25 199992.48  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#provide summary statistics of all float32 variables\n",
    "df.describe(include = ['float32']).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "CW1zb8NYx4MM",
    "outputId": "6ea1bf60-d89f-438d-c605-0dca60dc2889"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3617\n",
      "3202\n",
      "3181\n"
     ]
    }
   ],
   "source": [
    "#identify three categories of approximately equal size\n",
    "print(df[df.Balance==0].shape[0])\n",
    "print(df[df.Balance<120000].shape[0]-df[df.Balance==0].shape[0])\n",
    "print(df[df.Balance>120000].shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "SFPdxG2x3Nt3"
   },
   "outputs": [],
   "source": [
    "#identify a median balance\n",
    "median_balance = 120000\n",
    "\n",
    "#create bins\n",
    "bins = [-1,1,median_balance,np.inf]\n",
    "\n",
    "#create labels for each bin\n",
    "category = ['<$1','$1-'+str(median_balance)[0:3]+'K','$'+str(median_balance)[0:3]+'K']\n",
    "\n",
    "#cut all observations into one of three bins\n",
    "df['Balance'] = pd.cut(df['Balance'], bins, labels=category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "y_iq6OQi4RGS",
    "outputId": "c3fe00ee-8f5e-4416-8926-639f1735327c"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<$1        3617\n",
       "$1-120K    3202\n",
       "$120K      3181\n",
       "Name: Balance, dtype: int64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#confirm transformation\n",
    "df.Balance.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "M6cj6d9sdYPG"
   },
   "source": [
    "#### Observations\n",
    "- The distribution for *Balance* appears highly left skewed, with 3617 people having a balance of $0 and the remaining cohort having a normal distribution of balances centered at approximately 120,000 dollars.\n",
    "  - The information provided by this variable would be better served within the model by a categorical variable.\n",
    "  - Based on my interpretation, this new variable should have values for a zero balance, a value for a balance below median, and a value for above median of this normal distribution.\n",
    "- The distribution for \"EstimatedSalary\" appears relatively uniform between the minimum and maximum values of 12 dollars and 200,000 dollars, respectively."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EVsMCtsjZvQx"
   },
   "source": [
    "### Summary of integer variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "9ajjBmDSZ1-N",
    "outputId": "5afc2329-cdd5-4529-8a21-a03f215840d6"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1152x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1152x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1152x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7kAAAFzCAYAAAD7S847AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de7hdZX0v+u9PrtKqIEmFE4IRRap4FDWiom3Z4lZ0g1iLwN6I0AfBXbWVnrZb4XQfL1gv+9Etuo9YEXlARSHHS7mIF0BRu/FC0AiCRiNKSQpybbygQOQ9f6wRXKZZYSVkrjHnyOfzPOtZY7xjzDm/U4dj+c075pjVWgsAAAAMwYP6DgAAAACbi5ILAADAYCi5AAAADIaSCwAAwGAouQAAAAyGkgsAAMBgbN13gFGYN29eW7RoUd8xAAAAGIErr7zy1tba/PVtG2TJXbRoUZYuXdp3DAAAAEagqq6faZvLlQEAABgMJRcAAIDBUHIBAAAYDCUXAACAwRjkjafG3Wtf+9qsXr06u+66a99RgAmyxx575Ljjjus7BgDAWFNye3DTTTfll7/+df5lp536jgJMiO1Xreo7AgDARFBye3Lvttvmx699bd8xgAnxqPe8p+8IAAATwWdyAQAAGAwlFwAAgMFQcgEAABgMJRcAAIDBUHIBAAAYDCUXAACAwVByAQAAGAwlFwAAgMFQcgEAABgMJRcAAIDBUHIBAAAYDCUXAACAwdi67wBborvvvjvVWt8xAICB++AHP5gkOe6443pOAjB3lNwe3HvvvUouADBy1113Xd8RAOacy5UBAAAYDCUXAACAwVByAQAAGAwlFwAAgMFQcgEAABgMJRcAAIDBUHIBAAAYDCUXAACAwVByAQAAGAwlFwAAgMHYuu8AAAAAjI+DDz74vuULLrigxySbxkwuTJC66/Zsv+z1qbvv6DsKAACMJSUXJsg215+TrVZfm22uP6fvKAAADND0Wdz1rU8ClyvDhKi7bs82P700lZZtbrok9zzyiLRtd+o7FnNk21tuyXX33JMTTzyx7yjABLnuuuvy4Ac/uO8YAHNqMDO5VXV8VS2tqqW33HJL33Fgs9vm+nOSdu/USrvXbC4AAKzHYGZyW2unJTktSRYvXtx6jgOb1X2zuG3N1HpbYzZ3C3P3/PnZY/vt87a3va3vKMAEcfUHsCUazEwuDNnvzOKuZTYXAAD+HSUXJsBWP/v+fbO4a1Vbk61Wf6+nRAAADNG6Xxk0iV8hNJjLlWHIfrX4vX1HAACAiaDkAgAAcJ9JnL2dzuXKAAAADIaSCwAAwGAouQAAAAyGkgsAAMBgKLkAAAAMhpILAADAYCi5AAAADIaSCwAAwGAouQAAAAzG1n0H2BI96EEPyprW+o4BAAzcHnvs0XcEgDmn5PZg2223zV333tt3DABg4I477ri+IwDMOZcrAwAAMBhKLgAAAIOh5AIAADAYSi4AAACDoeQCAAAwGEouAAAAg6HkAgAAMBhKLgAAAIOh5AIAADAYSi4AAACDoeQCAAAwGFv3HWBL9aC7786j3vOevmMAE2L7VauSRz+67xgAAGNPye3BLrvsktWrV2fX7bfvOwowKR796Oyxxx59pwAAGHtKbg/eYwYXAABgJHwmFwAAgMFQcgEAABgMJRcAAIDBUHIBAAAYDCUXAACAwVByAQAAGIxqrfWdYbOrqluSXN93jvsxL8mtfYdgLDk2mIljgw1xfDATxwYzcWywIeN+fDyytTZ/fRsGWXInQVUtba0t7jsH48exwUwcG2yI44OZODaYiWODDZnk48PlygAAAAyGkgsAAMBgKLn9Oa3vAIwtxwYzcWyMgapqVfWuaet/W1Vv3IzPf3xVfb/7+WZVPXvatj+qqmuqallVPa6qftUtX5vkl1W1yX/Xq+onVTVvEx63qKr+y6a+LnPCuYOZODbYkIk9PpTcnrTWJvagYbQcG8zEsTE27krykk0phPenqg5K8sokz26t/WGS/5rkY1W1S7fLkUne1lrbJ8mvkvyoW35ikkry4nWeb+vNnXE9FiVRcseYcwczcWywIZN8fCi5ALBx1mTqX7f/et0NVXVmVR06bf0X3e/9q+rLVXVeVV1XVW+vqiO7mdqrq+rR3UNel+TvWmu3Jklr7VtJzkry6qp6RZLDkpxcVWdPf93W2poklyd5TFUdU1XnV9UXk1xaVQ+vqn+qqquq6utV9cQu085V9YVuZvj0TJXktTOz3532Hu6bqa6qx1TVJVX1nar6Vpf77Un+qJtR/uuq2rt7X8u619zzAf8nDgAbQckFgI33viRHVtXDNuIxT8rUzOzjkhyV5LGttX2TnJ7kL7t99k5y5TqPW5pk79ba6UnOz1QJPnL6DlW1Q5IDklzdDT0lyaGttT9J8qYk326tPTHJSUk+3O3zhiT/3FrbO8mnk+w+i/dwdpL3tdaelGS/JDcmeX2Sr7bW9mmtvbt7j+/pZpgXJ1k5i+cFgM1GyQWAjdRa+1mmyuJfbcTDrmit3dhauyvJj5J8oRu/OlOX/G6KR1fVsiT/O8lnWmuf7cYvbq3d3i0/O8lHutxfTLJzVT00yR8n+Wg3/pkkd2zoharqIUkWtNY+3T3m1621O9ez69eSnFRVr8vUdxj+ahPfGwBsEiUXADbNKUmOTfJ708bWpPvb2t0Eattp2+6atnzvtPV7k6z97Oy1SZ66zus8Nck1M2T4UTeD+uTW2hunjf9ylu9hfe57D53tN+bBrbWPJXlRpj4zfFFVPecBZAGAjabkAsAm6GZKl2Sq6K71k/y2pL4oyTYb+bT/I8k7qmrnJKmqfZIck+TUBxD1q5m6YVWqav8kt3Yz0V9Jd8OoqnpBkp26/X+a5A+6z+xul+SgJGmt/TzJyqp6cfeY7brLpH+e5CFrX6yq9khyXWvtvUnOy9RNsQBgzszFXRcBYKjeleQ109Y/mOS8qvpOks9lI2dUW2vnV9WCJJdXVctUgXxZa+3GB5DxjUnOqKqrktyZ5Ohu/E1JPl5V12TqplX/0mW4p6renOSbSVYl+f605zoqyQe67fckeWmSq5L8pnvPZybZLslRVXVPkpuSvPUBZAeAjVattb4zAAAAwGbhcmUAAAAGQ8kFAABgMJRcAAAABkPJBQAAYDCUXAAAAAZDyQUAAGAwlFwAAAAGQ8kFAABgMJRcAAAABkPJBQAAYDCUXAAAAAZDyQUAAGAwlFwAAAAGQ8kFAABgMJRcAAAABkPJBQAAYDC27jvAKMybN68tWrSo7xgzWr58eZJkr7326jkJk2b1r1cnSR62/cN6TgIAAP258sorb22tzV/ftkGW3EWLFmXp0qV9x5jR/vvvnyS57LLLes0BAAAwiarq+pm2uVwZJsjyW5dn+a3L+44BAABja5AzuTBUr7zwlUmSy465rN8gAAAwpszkAgAAMBhKLgAAAIOh5AIAADAYSi4AAACD4cZTMEH+/o//vu8IAAAw1pRcmCDP3eO5fUcAAICx5nJlmCDLblqWZTct6zsGAACMrZGW3Kr6SVVdXVXLqmppN/bwqrq4qn7Y/d6pG6+qem9Vraiqq6rqKdOe5+hu/x9W1dGjzDxXvva1r6eqRvqzYOHufb9NNrMTPndCTvjcCX3HAACAsTUXlyv/h9bardPWX5/k0tba26vq9d3665K8IMme3c/Tk7w/ydOr6uFJ3pBkcZKW5MqqOr+1dsccZB+Zu+++K4d/4PKRvsa5r9xvpM8PAAAwbvq4XPmQJGd1y2clefG08Q+3KV9PsmNV7Zrk+Ukubq3d3hXbi5McONehAQAAGH+jLrktyReq6sqqOr4be0Rr7cZu+aYkj+iWFyS5YdpjV3ZjM40DAADA7xj15crPbq2tqqo/SHJxVX1/+sbWWquqtjleqCvRxyfJ7rv7LCoAAMCWaKQlt7W2qvt9c1V9Osm+SX5aVbu21m7sLke+udt9VZKF0x6+Wze2Ksn+64xftp7XOi3JaUmyePHizVKcYdy89YC39h0BAADG2sguV66q36uqh6xdTvK8JN9Ncn6StXdIPjrJed3y+Ule3t1l+RlJVneXNX8+yfOqaqfuTszP68Zgi7Pfwv2y30I3FAMAgJmMcib3EUk+XVVrX+djrbXPVdUVSZZU1bFJrk9yWLf/RUlemGRFkjuT/HmStNZur6qTk1zR7ffm1trtI8wNY+vyG6buyK3oAgDA+o2s5LbWrkvypPWM35bkgPWMtySvnuG5zkhyxubOCJPmpEtPSpJcdsxl/QYBAIAx1cdXCAEAAMBIKLkAAAAMhpILAADAYCi5AAAADMZIvycX2LxOOfCUviMAAMBYU3Jhguyzyz59RwAAgLHmcmWYIJdcd0kuue6SvmMAAMDYMpMLE+QtX3lLkuS5ezy35yQAADCezOQCAAAwGEouAAAAg6HkAgAAMBhKLgAAAIPhxlMwQT5w0Af6jkCPFizcPf+68oaRvsb/sdvCrLrhX0b6GgAAo6TkwgTZa95efUegR/+68oYc/oHLR/oa575yv5E+PwDAqLlcGSbIBcsvyAXLL+g7BgAAjC0zuTBB3vW1dyVJDt7r4J6TAADAeDKTCwAAwGAouQAAAAyGkgsAAMBgKLkAAAAMhhtPwQT5yJ9+pO8IAAAw1pRcmCALH7aw7wgAADDWXK4ME+Tc756bc797bt8xAABgbJnJhQny/qXvT5Ic/oTDe04CAADjyUwuAAAAg6HkAgAAMBhKLgAAAIOh5AIAADAYbjwFE+QTh32i7wgAADDWlFyYIPN2mNd3BAAAGGsuV4YJcuayM3PmsjP7jgEAAGNLyYUJouQCAMCGKbkAAAAMxshLblVtVVXfrqoLu/VHVdU3qmpFVZ1bVdt249t16yu67YumPceJ3fjyqnr+qDMDAAAwmeZiJve1Sb43bf0dSd7dWntMkjuSHNuNH5vkjm783d1+qarHJzkiyd5JDkxyalVtNQe5AQAAmDAjLblVtVuS/5Tk9G69kjwnydrvQTkryYu75UO69XTbD+j2PyTJOa21u1prP06yIsm+o8wNAADAZBr1VwidkuS/JXlIt75zkn9rra3p1lcmWdAtL0hyQ5K01tZU1epu/wVJvj7tOac/5j5VdXyS45Nk991337zvAsbERUde1HcEAAAYayObya2qg5Lc3Fq7clSvMV1r7bTW2uLW2uL58+fPxUvCnNthmx2ywzY79B0DAADG1ihncp+V5EVV9cIk2yd5aJL3JNmxqrbuZnN3S7Kq239VkoVJVlbV1kkeluS2aeNrTX8MbFFOveLUJMmrnvaqnpMAAMB4GtlMbmvtxNbabq21RZm6cdQXW2tHJvlSkkO73Y5Ocl63fH63nm77F1trrRs/orv78qOS7Jnkm6PKDeNsyTVLsuSaJX3HAACAsTXqz+Suz+uSnFNVb0ny7SQf6sY/lOQjVbUiye2ZKsZprV1TVUuSXJtkTZJXt9Z+M/exAQAAGHdzUnJba5cluaxbvi7ruTtya+3XSV46w+P/Ick/jC4hAAAAQzAX35MLAAAAc0LJBQAAYDD6+EwusIkuO+ayviMAAMBYM5MLAADAYCi5MEHeefk7887L39l3DAAAGFtKLkyQC39wYS78wYV9xwAAgLGl5AIAADAYSi4AAACDoeQCAAAwGL5CCCbIg7d5cN8RAABgrCm5MEE+e+Rn+44AAABjzeXKAAAADIaSCxPk5C+fnJO/fHLfMQAAYGwpuTBBLv3xpbn0x5f2HQMAAMaWkgsAAMBgKLkAAAAMhpILAADAYPgKIZggO++wc98RAABgrCm5MEE+edgn+44AAABjzeXKAAAADIaSCxPkxEtOzImXnNh3DAAAGFsuV4YJ8rWVX+s7AgAAjDUzuQAAAAyGkgsAAMBgKLkAAAAMhs/kwgTZ7aG79R0BAADGmpILE+SjL/lo3xEAAGCsuVwZAACAwVByYYKc8LkTcsLnTug7BgAAjC2XK8MEWXbTsr4jAADAWDOTCwAAwGAouQAAAAzGyEpuVW1fVd+squ9U1TVV9aZu/FFV9Y2qWlFV51bVtt34dt36im77omnPdWI3vryqnj+qzAAAAEy2Uc7k3pXkOa21JyXZJ8mBVfWMJO9I8u7W2mOS3JHk2G7/Y5Pc0Y2/u9svVfX4JEck2TvJgUlOraqtRpgbxtZjd35sHrvzY/uOAQAAY2tkJbdN+UW3uk3305I8J8knuvGzkry4Wz6kW0+3/YCqqm78nNbaXa21HydZkWTfUeWGcXbawafltINP6zsGAACMrZF+JreqtqqqZUluTnJxkh8l+bfW2ppul5VJFnTLC5LckCTd9tVJdp4+vp7HAAAAwH1GWnJba79pre2TZLdMzb7+4aheq6qOr6qlVbX0lltuGdXLQK+Ov+D4HH/B8X3HAACAsTWrkltVz5rN2Exaa/+W5EtJnplkx6pa+/28uyVZ1S2vSrKwe+6tkzwsyW3Tx9fzmOmvcVprbXFrbfH8+fNnGw0myg9u+0F+cNsP+o4BAABja7Yzuf9rlmP3qar5VbVjt/zgJP8xyfcyVXYP7XY7Osl53fL53Xq67V9srbVu/Iju7suPSrJnkm/OMjcAAABbkK03tLGqnplkvyTzq+r/mrbpoUnu7w7HuyY5q7sT8oOSLGmtXVhV1yY5p6rekuTbST7U7f+hJB+pqhVJbs/UHZXTWrumqpYkuTbJmiSvbq39ZmPeJAAAAFuGDZbcJNsm+f1uv4dMG/9Zfjsbu16ttauSPHk949dlPXdHbq39OslLZ3iuf0jyD/eTFQAAgC3cBktua+3LSb5cVWe21q6fo0zADPbZZZ++IwAAwFi7v5nctbarqtOSLJr+mNbac0YRCli/Uw48pe8IAAAw1mZbcv+/JP+Y5PQkPg8LAADAWJptyV3TWnv/SJMA9+tln3pZkuSjL/loz0kAAGA8zbbkXlBVr0ry6SR3rR1srd0+klTAeq382cq+IwAAwFibbcld+/21fzdtrCXZY/PGAQAAgE03q5LbWnvUqIMAAADAAzWrkltVL1/feGvtw5s3DgAAAGy62V6u/LRpy9snOSDJt5IouTCHnrnbM/uOAAAAY222lyv/5fT1qtoxyTkjSQTM6G3PfVvfEQAAYKw9aBMf98skPqcLAADAWJntZ3IvyNTdlJNkqySPS7JkVKGA9fuzJX+WJPnkYZ/sOQkAAIyn2X4m953Tltckub615gs7YY7ddudtfUcAAICxNqvLlVtrX07y/SQPSbJTkrtHGQoAAAA2xaxKblUdluSbSV6a5LAk36iqQ0cZDAAAADbWbC9X/r+TPK21dnOSVNX8JJck+cSoggEAAMDGmm3JfdDagtu5LZt+Z2ZgEx3wqAP6jgAAAGNttiX3c1X1+SQf79YPT3LRaCIBM/nvf/Lf+44AAABjbYMlt6oek+QRrbW/q6qXJHl2t+lrSc4edTgAAADYGPc3k3tKkhOTpLX2qSSfSpKq+j+7bQePNB3wO15w9guSJJ898rM9JwEAgPF0fyX3Ea21q9cdbK1dXVWLRpIImNGv7vlV3xEAAGCs3d/No3bcwLYHb84gAAAA8EDdX8ldWlXHrTtYVa9IcuVoIgEAAMCmub/LlU9I8umqOjK/LbWLk2yb5E9HGQwAAAA21gZLbmvtp0n2q6r/kOQJ3fBnWmtfHHky4N856LEH9R0BAADG2qy+J7e19qUkXxpxFuB+/O1+f9t3BAAAGGv395lcAAAAmBhKLkyQ/c/cP/ufuX/fMQAAYGwpuQAAAAyGkgsAAMBgKLkAAAAMhpILAADAYMzqK4SA8XDY3of1HQEAAMbayGZyq2phVX2pqq6tqmuq6rXd+MOr6uKq+mH3e6duvKrqvVW1oqquqqqnTHuuo7v9f1hVR48qM4y7Vz3tVXnV017VdwwAABhbo7xceU2Sv2mtPT7JM5K8uqoen+T1SS5tre2Z5NJuPUlekGTP7uf4JO9PpkpxkjckeXqSfZO8YW0xhi3NnffcmTvvubPvGAAAMLZGVnJbaze21r7VLf88yfeSLEhySJKzut3OSvLibvmQJB9uU76eZMeq2jXJ85Nc3Fq7vbV2R5KLkxw4qtwwzl549gvzwrNf2HcMAAAYW3Ny46mqWpTkyUm+keQRrbUbu003JXlEt7wgyQ3THrayG5tpfN3XOL6qllbV0ltuuWWz5gcAAGAyjLzkVtXvJ/lkkhNaaz+bvq211pK0zfE6rbXTWmuLW2uL58+fvzmeEgAAgAkz0pJbVdtkquCe3Vr7VDf80+4y5HS/b+7GVyVZOO3hu3VjM40DAADA7xjl3ZUryYeSfK+19j+nbTo/ydo7JB+d5Lxp4y/v7rL8jCSru8uaP5/keVW1U3fDqed1YwAAAPA7Rvk9uc9KclSSq6tqWTd2UpK3J1lSVccmuT7J2i/+vCjJC5OsSHJnkj9Pktba7VV1cpIruv3e3Fq7fYS5YWwds88xfUcAAICxNrKS21r75yQ1w+YD1rN/S/LqGZ7rjCRnbL50MJmUXAAA2LA5ubsysHnceuetufXOW/uOAQAAY2uUlysDm9mhSw5Nklx2zGX9BgEAgDFlJhcAAIDBUHIBAAAYDCUXAACAwVByAQAAGAw3noIJ8heL/6LvCAAAMNaUXJgghz/h8L4jAADAWHO5MkyQG1bfkBtW39B3DAAAGFtmcmGCHPXpo5L4nlwAAJiJmVwAAAAGQ8kFAABgMJRcAAAABkPJBQAAYDDceAomyN8882/6jgAAAGNNyYUJcvBeB/cdAQAAxprLlWGCLL91eZbfurzvGAAAMLbM5MIEeeWFr0zie3IBAGAmZnIBAAAYDCUXAACAwVByAQAAGAwlFwAAgMFw4ymYIH//x3/fdwQAABhrSi5MkOfu8dy+IwAAwFhzuTJMkGU3Lcuym5b1HQMAAMaWmVyYICd87oQkvicXAABmYiYXAACAwVByAQAAGAwlFwAAgMFQcgEAABgMN56CCfLWA97adwQAABhrSi5MkP0W7td3BAAAGGsuV4YJcvkNl+fyGy7vOwYAAIytkZXcqjqjqm6uqu9OG3t4VV1cVT/sfu/UjVdVvbeqVlTVVVX1lGmPObrb/4dVdfSo8sIkOOnSk3LSpSf1HQMAAMbWKGdyz0xy4Dpjr09yaWttzySXdutJ8oIke3Y/xyd5fzJVipO8IcnTk+yb5A1rizEAAACsa2Qlt7X2lSS3rzN8SJKzuuWzkrx42viH25SvJ9mxqnZN8vwkF7fWbm+t3ZHk4vz74gwAAABJ5v4zuY9ord3YLd+U5BHd8oIkN0zbb2U3NtP4v1NVx1fV0qpaesstt2ze1AAAAEyE3m481VprSdpmfL7TWmuLW2uL58+fv7meFgAAgAky118h9NOq2rW1dmN3OfLN3fiqJAun7bdbN7Yqyf7rjF82BzlhLJ1y4Cl9RwAAgLE21zO55ydZe4fko5OcN2385d1dlp+RZHV3WfPnkzyvqnbqbjj1vG4Mtkj77LJP9tlln75jAADA2BrZTG5VfTxTs7Dzqmplpu6S/PYkS6rq2CTXJzms2/2iJC9MsiLJnUn+PElaa7dX1clJruj2e3Nrbd2bWcEW45LrLkmSPHeP5/acBAAAxtPISm5r7T/PsOmA9ezbkrx6huc5I8kZmzEaTKy3fOUtSZRcAACYSW83ngIAAIDNTckFAABgMJRcAAAABkPJBQAAYDDm+ntygQfgAwd9oO8IAAAw1pRcmCB7zdur7wgAADDWXK4ME+SC5RfkguUX9B0DAADGlplcmCDv+tq7kiQH73Vwz0kAAGA8mckFAABgMJRcAAAABkPJBQAAYDCUXAAAAAbDjadggnzkTz/SdwQAABhrSi5MkIUPW9h3BAAAGGsuV4YJcu53z8253z237xgAADC2zOTCBHn/0vcnSQ5/wuE9JwEAgPFkJhcAAIDBUHIBAAAYDCUXAACAwVByAQAAGAw3noIJ8onDPtF3BAAAGGtKLkyQeTvM6zsCAACMNZcrwwQ5c9mZOXPZmX3HAACAsaXkwgRRcgEAYMOUXAAAAAZDyQUAGLAFC3dPVY30Z8HC3ft+mwD3ceMpAIAB+9eVN+TwD1w+0tc495X7jfT5ATaGmVwAAAAGw0wuTJCLjryo7wgAADDWlFyYIDtss0PfEQAAYKy5XBkmyKlXnJpTrzi17xgAADC2lFyYIEuuWZIl1yzpOwYAAIytiSm5VXVgVS2vqhVV9fq+8wAAADB+JqLkVtVWSd6X5AVJHp/kP1fV4/tNBQAAMPd8//WGTcqNp/ZNsqK1dl2SVNU5SQ5Jcm2vqQAAAOaY77/esImYyU2yIMkN09ZXdmMAAABwn2qt9Z3hflXVoUkObK29ols/KsnTW2uvmbbP8UmO71b3SrJ8zoNunHlJbu07BGPJscFMHBtsiOODmTg2mIljgw0Z9+Pjka21+evbMCmXK69KsnDa+m7d2H1aa6clOW0uQz0QVbW0tba47xyMH8cGM3FssCGOD2bi2GAmjg02ZJKPj0m5XPmKJHtW1aOqatskRyQ5v+dMAAAAjJmJmMltra2pqtck+XySrZKc0Vq7pudYAAAAjJmJKLlJ0lq7KMlFfefYjCbm0mrmnGODmTg22BDHBzNxbDATxwYbMrHHx0TceAoAAABmY1I+kwsAAAD3S8kdoao6o6purqrvzrC9quq9VbWiqq6qqqfMdUb6MYtjY/+qWl1Vy7qf/2euM9KPqlpYVV+qqmur6pqqeu169nHu2ALN8thw7thCVdX2VfXNqvpOd3y8aT37bFdV53bnjm9U1aK5T8pcm+WxcUxV3TLt3PGKPrLSj6raqqq+XVUXrmfbRJ43JuYzuRPqzCT/b5IPz7D9BUn27H6enuT93W+G78xs+NhIkq+21g6amziMkTVJ/qa19q2qekiSK6vq4tbatdP2ce7YMs3m2EicO7ZUdyV5TmvtF1W1TZJ/rqrPtta+Pm2fY5Pc0Vp7TFUdkeQdSQ7vIyxzajbHRpKc21p7TQ/56N9rk3wvyUPXs20izxtmckeotfaVJLdvYJdDkny4Tfl6kh2rate5SUefZnFssIVqrd3YWvtWt/zzTP3RWbDObs4dW6BZHhtsobrzwS+61W26n6/b4FMAAAZ5SURBVHVvvHJIkrO65U8kOaCqao4i0pNZHhtsoapqtyT/KcnpM+wykecNJbdfC5LcMG19ZfwfFn7rmd2lRZ+tqr37DsPc6y4JenKSb6yzybljC7eBYyNx7thidZccLktyc5KLW2sznjtaa2uSrE6y89ympA+zODaS5M+6j8B8oqoWznFE+nNKkv+W5N4Ztk/keUPJhfH0rSSPbK09Kcn/SvJPPedhjlXV7yf5ZJITWms/6zsP4+N+jg3nji1Ya+03rbV9kuyWZN+qekLfmRgPszg2LkiyqLX2xCQX57czdwxYVR2U5ObW2pV9Z9nclNx+rUoy/V/KduvG2MK11n629tKi7juit6mqeT3HYo50n5n6ZJKzW2ufWs8uzh1bqPs7Npw7SJLW2r8l+VKSA9fZdN+5o6q2TvKwJLfNbTr6NNOx0Vq7rbV2V7d6epKnznU2evGsJC+qqp8kOSfJc6rqo+vsM5HnDSW3X+cneXl3p9RnJFndWrux71D0r6p2Wft5h6raN1P/Wx37EwoPXPff+4eSfK+19j9n2M25Yws0m2PDuWPLVVXzq2rHbvnBSf5jku+vs9v5SY7ulg9N8sXWms9mDtxsjo117uvwokx95p+Ba62d2FrbrbW2KMkRmTonvGyd3SbyvOHuyiNUVR9Psn+SeVW1MskbMvVh/7TW/jHJRUlemGRFkjuT/Hk/SZlrszg2Dk3yF1W1JsmvkhwxCScUNotnJTkqydXd56eS5KQkuyfOHVu42Rwbzh1brl2TnFVVW2XqHzeWtNYurKo3J1naWjs/U/9I8pGqWpGpmx8e0V9c5tBsjo2/qqoXZeou7rcnOaa3tPRuCOeN8rcPAACAoXC5MgAAAIOh5AIAADAYSi4AAACDoeQCAAAwGEouAAAAg6HkAsBGqKpWVe+atv63VfXGzfj8x1fV97ufb1bVs6dt+6OquqaqllXV46rqV93ytVX1j1W1yX/Xq+onVTVvEx63qKr+y6a+LgBsbkouAGycu5K8ZFMK4f2pqoOSvDLJs1trf5jkvyb5WFXt0u1yZJK3tdb2ydT34P6oW35ikscnefE6z7f15s64HouSKLkAjA0lFwA2zpokpyX563U3VNWZVXXotPVfdL/3r6ovV9V5VXVdVb29qo7sZmqvrqpHdw95XZK/a63dmiSttW8lOSvJq6vqFUkOS3JyVZ09/XVba2uSXJ7kMVV1TFWdX1VfTHJpVT28qv6pqq6qqq9X1RO7TDtX1Re6meHTk1Q3vqiqvjvtPdw3U11Vj6mqS6rqO1X1rS7325P8UTej/NdVtXf3vpZ1r7nnA/5PHAA2gpILABvvfUmOrKqHbcRjnpSpmdnHJTkqyWNba/smOT3JX3b77J3kynUetzTJ3q2105Ocn6kSfOT0HapqhyQHJLm6G3pKkkNba3+S5E1Jvt1ae2KSk5J8uNvnDUn+ubW2d5JPJ9l9Fu/h7CTva609Kcl+SW5M8vokX22t7dNae3f3Ht/TzTAvTrJyFs8LAJuNkgsAG6m19rNMlcW/2oiHXdFau7G1dleSHyX5Qjd+daYu+d0Uj66qZUn+d5LPtNY+241f3Fq7vVt+dpKPdLm/mGTnqnpokj9O8tFu/DNJ7tjQC1XVQ5IsaK19unvMr1trd65n168lOamqXpfkka21X23iewOATaLkAsCmOSXJsUl+b9rYmnR/W7ubQG07bdtd05bvnbZ+b5K1n529NslT13mdpya5ZoYMP+pmUJ/cWnvjtPFfzvI9rM9976Gz/cY8uLX2sSQvytRnhi+qquc8gCwAsNGUXADYBN1M6ZJMFd21fpLfltQXJdlmI5/2fyR5R1XtnCRVtU+SY5Kc+gCifjVTN6xKVe2f5NZuJvor6W4YVVUvSLJTt/9Pk/xB95nd7ZIclCSttZ8nWVlVL+4es113mfTPkzxk7YtV1R5JrmutvTfJeZm6KRYAzJm5uOsiAAzVu5K8Ztr6B5OcV1XfSfK5bOSMamvt/KpakOTyqmqZKpAva63d+AAyvjHJGVV1VZI7kxzdjb8pycer6ppM3bTqX7oM91TVm5N8M8mqJN+f9lxHJflAt/2eJC9NclWS33Tv+cwk2yU5qqruSXJTkrc+gOwAsNGqtdZ3BgAAANgsXK4MAADAYCi5AAAADIaSCwAAwGAouQAAAAyGkgsAAMBgKLkAAAAMhpILAADAYCi5AAAADMb/D4evvmIO2M69AAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 1152x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1152x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot all int32 variables\n",
    "for col in df.columns[df.dtypes == 'int32']:\n",
    "    histogram_boxplot(df, col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "Xn6lGWlxZ1-L",
    "outputId": "3daea69a-979f-4795-ef94-b1db5e7a0cab"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-62803052-f2a4-43a7-9031-e0e570aaf8f9\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>CreditScore</th>\n",
       "      <td>10000.00</td>\n",
       "      <td>650.53</td>\n",
       "      <td>96.65</td>\n",
       "      <td>350.00</td>\n",
       "      <td>584.00</td>\n",
       "      <td>652.00</td>\n",
       "      <td>718.00</td>\n",
       "      <td>850.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Age</th>\n",
       "      <td>10000.00</td>\n",
       "      <td>38.92</td>\n",
       "      <td>10.49</td>\n",
       "      <td>18.00</td>\n",
       "      <td>32.00</td>\n",
       "      <td>37.00</td>\n",
       "      <td>44.00</td>\n",
       "      <td>92.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Tenure</th>\n",
       "      <td>10000.00</td>\n",
       "      <td>5.01</td>\n",
       "      <td>2.89</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>10.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NumOfProducts</th>\n",
       "      <td>10000.00</td>\n",
       "      <td>1.53</td>\n",
       "      <td>0.58</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>4.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Exited</th>\n",
       "      <td>10000.00</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-62803052-f2a4-43a7-9031-e0e570aaf8f9')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-62803052-f2a4-43a7-9031-e0e570aaf8f9 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-62803052-f2a4-43a7-9031-e0e570aaf8f9');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "                 count   mean   std    min    25%    50%    75%    max\n",
       "CreditScore   10000.00 650.53 96.65 350.00 584.00 652.00 718.00 850.00\n",
       "Age           10000.00  38.92 10.49  18.00  32.00  37.00  44.00  92.00\n",
       "Tenure        10000.00   5.01  2.89   0.00   3.00   5.00   7.00  10.00\n",
       "NumOfProducts 10000.00   1.53  0.58   1.00   1.00   1.00   2.00   4.00\n",
       "Exited        10000.00   0.20  0.40   0.00   0.00   0.00   0.00   1.00"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#provide summary statistics of all int32 variables\n",
    "df.describe(include = ['int32']).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "fmWBLHBOAneI",
    "outputId": "eae5f6a5-3e38-4eca-9324-8ef3fb547a0a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "850    233\n",
      "678     63\n",
      "655     54\n",
      "705     53\n",
      "667     53\n",
      "      ... \n",
      "404      1\n",
      "351      1\n",
      "365      1\n",
      "417      1\n",
      "419      1\n",
      "Name: CreditScore, Length: 460, dtype: int64\n",
      " \n",
      "37    478\n",
      "38    477\n",
      "35    474\n",
      "36    456\n",
      "34    447\n",
      "     ... \n",
      "92      2\n",
      "82      1\n",
      "88      1\n",
      "85      1\n",
      "83      1\n",
      "Name: Age, Length: 70, dtype: int64\n",
      " \n",
      "2     1048\n",
      "1     1035\n",
      "7     1028\n",
      "8     1025\n",
      "5     1012\n",
      "3     1009\n",
      "4      989\n",
      "9      984\n",
      "6      967\n",
      "10     490\n",
      "0      413\n",
      "Name: Tenure, dtype: int64\n",
      " \n",
      "1    5084\n",
      "2    4590\n",
      "3     266\n",
      "4      60\n",
      "Name: NumOfProducts, dtype: int64\n",
      " \n",
      "0    7963\n",
      "1    2037\n",
      "Name: Exited, dtype: int64\n",
      " \n"
     ]
    }
   ],
   "source": [
    "#provide counts for all distinct values within each int32 variable\n",
    "for col in df.columns[df.dtypes == 'int32']:\n",
    "    print(df[col].value_counts())\n",
    "    print(\" \")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "zAGrxieqdURk"
   },
   "source": [
    "#### Observations\n",
    "- The distribution of *CreditScore* appears relatively normal, with a center with a mean credit score of 650 and an upper limit at 850.\n",
    "- The distribution for *Age* appears right skewed, with a lower limit of 18, a mean age of 39, and a maximum age of 92.\n",
    "- The distribution for *Tenure* appears uniform with, ~1000 people each between 1 and 9 years and half as many in the cohorts for 1 and 10 years with the bank.\n",
    "- Only 326 of 10,000 people sampled have 3 or more products with the bank.\n",
    "\n",
    "**Dependent variable**\n",
    "- From the sample of 10,000 people, ~20% of customers subsequently left the bank.\n",
    "- Further analysis on this variable is included below as part of the bivariate analysis."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "jbsBq1J6ctu-"
   },
   "source": [
    "### Summary of categorical variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "tv95qrRIdFJC",
    "outputId": "7b12c026-18f3-40ee-d14d-603615bab63e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 288x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARsAAAE9CAYAAAA72MYKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de5gX5X338fcnKMj5EFalLMJGIShEUTdKW8REEzlcUZR6PoRYGlTgqdjQJ2pVCAa1jZVoDTTE2OAhoBYPmzx4QEU0PJ7AIgg0usAiIAIG5ajo4rd//Ga3P3EXFvjNLEs+r+uaa2fuuWfmnrngs/ccVxGBmVnavlLfDTCzPw8OGzPLhMPGzDLhsDGzTDhszCwTDhszy8RB9d2ANLRv3z66dOlS380w+7M0b968DyKiaOfyAzJsunTpwty5c+u7GWZ/liStqKncp1FmlgmHTR1deumldOjQgVatWtGtWzfuueee6nnbtm1j+PDhtG/fntatW9O3b98a17F9+3aGDh1K586dadmyJb169eLJJ5+snr9y5Up69+5Nu3bt+NGPfvSFZQcMGODemjVoDps6uu6666ioqGDTpk2UlZVxww03MG/ePACGDRvGhg0bWLJkCRs2bGDChAk1rqOyspJOnToxe/ZsNm7cyE9/+lPOP/98KioqALj11lsZMmQIy5cv5/HHH68Ol4ceeoiSkhJKS0sz2VezNByQ12zS0KNHj+pxSUhi6dKlNG/enLKyMlatWkWrVq0AOPHEE2tcR/PmzRk7dmz19Pe+9z1KSkqYN28eXbp0Yfny5Vx99dW0bt2ab37zmyxbtoxu3bpx2223MWvWrFT3zyxt7tnsgeHDh9OsWTO6d+9Ohw4dGDhwIK+99hqdO3dmzJgxtG/fnm984xtMnz69Tutbu3Ytb7/9dnWQ9ezZk5kzZ/LRRx8xb948evTowY033sioUaNo06ZNmrtmljqHzR6YOHEimzdv5qWXXmLw4ME0adKEVatW8dZbb9G6dWvee+897r77boYMGcKSJUt2ua7PPvuMSy65hCFDhtC9e3cgd6r20ksvceqppzJ8+HA+/fRTFixYwJlnnsnFF19M3759ufvuu7PYVbOC04H4iYnS0tJI+2LqlVdeyTHHHMOOHTv48Y9/zLZt2zjooNxZ6Zlnnsl3vvMdrr766hqX/fzzz7n44ovZtGkTTzzxBAcffHCNdfr27cu///u/c99999GmTRtGjx7NCSecwCOPPMLRRx+d6v6Z7S1J8yLiSxcY3bPZS5WVlSxdupRjjz32S/Mk1bpcRDB06FDWrl3L9OnTawwagMmTJ9O7d2969uzJwoULKS0tpXHjxnzjG99g4cKFBdsPs6w4bOpg3bp1TJs2jS1btrBjxw6efvpppk6dyumnn07fvn054ogjuPXWW6msrGTOnDnMmjWLfv361biuq666iiVLlvC73/2Opk2b1rq9X/ziF9UXk0tKSpg1axZbtmxh7ty5fO1rX0trV83SExEH3HDiiSdGIa1bty769u0brVu3jpYtW0bPnj1j8uTJ1fPfeuut6N27dzRr1iyOPvroePTRR6vnjR8/Pvr37x8RERUVFQFEkyZNonnz5tXDAw888IXtXXbZZfHwww9XT7/77rtx0kknRZs2beKaa64p6L6ZFRowN2r4f+lrNmZWULVds/mzfM7mjAvG1XcTMvPMQzfVdxPMAF+zMbOMOGzMLBMOGzPLhMPGzDKRWthIOkTSa5LelLRI0k+S8t9IWi5pfjL0Ssol6S5J5ZIWSDohb11DJL2TDEPSarOZpSfNu1HbgdMiYoukg4E/SKr6eMs/RsR/7lR/ANA1GU4GJgEnS2oHjAFKgQDmSSqLiA9TbLuZFVhqPZvk+Z4tyeTBybCrh3oGAfcly70CtJHUAegHzIyIDUnAzAT6p9VuM0tHqtdsJDWSNB9YRy4wXk1mjU9OlSZIapKUdQRW5i2+KimrrXznbQ2TNFfS3PXr1xd8X8xs36QaNhGxIyJ6AcXASZJ6AtcB3YFvAu2AHxdoW5MjojQiSouKvvRhdzOrZ5ncjYqIj4BZQP+IWJOcKm0H/gM4Kam2GuiUt1hxUlZbuZk1IGnejSqS1CYZbwp8F/jv5DoMyn2H4WzgrWSRMuD7yV2p3sDGiFgDPA2cIamtpLbAGUmZmTUgad6N6gBMkdSIXKg9HBG/l/S8pCJAwHzgyqT+DGAgUA5sAy4HiIgNkm4GXk/qjYuIDSm228xSkFrYRMQC4Pgayk+rpX4AI2qZdy9wb0EbaGaZ8hPEZpYJh42ZZcJhY2aZcNiYWSYcNmaWCYeNFdSll15Khw4daNWqFd26deOee+4B4JVXXuG73/0u7dq1o6ioiPPOO481a9bUup4NGzZwzjnn0Lx5czp37sxvf/vb6nlvvvkmPXr0oH379txxxx3V5Z999hknn3wyK1eurGmVVs8cNlZQ1113HRUVFWzatImysjJuuOEG5s2bx4cffsiwYcOoqKhgxYoVtGzZkssvv7zW9YwYMYLGjRuzdu1aHnzwQa666ioWLVpUvY3bb7+dN998k/Hjx/P+++8DcMcdd/A3f/M3dOrUqdb1Wv1x2FhB9ejRgyZNcu/WSkISS5cuZcCAAZx33nm0atWKZs2aMXLkSObMmVPjOrZu3cr06dO5+eabadGiBX369OGss87i/vvvB2D58uWcdtppdOzYka5du/Luu++yYsUKpk+fzjXXXJPZvu6L2nqAn376Keeeey5dunRBEi+88MIu11NRUcHAgQNp27Ythx9+OCNHjqSyshKAjRs30q9fP9q0acMll1zCjh07qpcbNmwYjz76aGr7VxOHjRXc8OHDadasGd27d6dDhw4MHDjwS3VefPFFevToUePyb7/9NgcddBDdunWrLjvuuOOqezY9e/bkmWeeYdWqVVRUVHDkkUdy9dVX87Of/azWvzC6v6mtBwjQp08fHnjgAQ4//PDdrmf48OEceuihrFmzhvnz5zN79mwmTpwIwC9/+UuOP/541q5dS0VFBY899hgAL7/8Mu+99x6DBw9Obwdr4LCxgps4cSKbN2/mpZdeYvDgwdU9nSoLFixg3Lhx/OxnP6tx+S1bttCqVasvlLVu3ZrNmzcDcPvttzNp0iTOOussJkyYwJw5c2jZsiUlJSUMGjSIU089lUceeSSdnSuQ2nqAjRs3ZtSoUfTp04dGjRrtdj3Lly/n/PPP55BDDuHwww+nf//+1aG8fPlyvv3tb9OkSRNOOeUUli1bxo4dO7jmmmu46667Ut2/mjhsLBWNGjWiT58+rFq1ikmTJlWXl5eXM2DAAO68805OOeWUGpdt0aIFmzZt+kLZpk2baNmyJQCdO3dmxowZvPHGGwwaNIgbb7yR22+/ndGjR3PBBRdQVlbGP/zDP7Bhw/79Cl1deoC7M2rUKKZNm8a2bdtYvXo1Tz75JP37574t17NnT5599lk+/vhjXnrpJXr06MFdd93FgAED6uVPODtsLFWVlZUsXboUgBUrVvCd73yHG2+8kcsuu6zWZbp160ZlZSXvvPNOdVnVHaidjRs3jh/+8IccdthhLFy4kNLSUlq3bk1xcTHl5eWF36EC2l0PsC769u3LokWLaNWqFcXFxZSWlnL22WcDMHToUDZu3MjJJ5/MKaecwnHHHcf999/PqFGjuPLKK+nbty833HBDoXerVg4bK5h169Yxbdo0tmzZwo4dO3j66aeZOnUqp59+OqtXr+a0005j5MiRXHnllbtcT/PmzRk8eDA33XQTW7duZc6cOTzxxBNfCqjFixfzwgsvcNVVVwFQUlLC888/z9q1a3nnnXc44ogjUtvXQqmtB1gXn3/+Of3792fw4MFs3bqVDz74gA8//JAf/zj3PbpDDjmEyZMns2DBAm677TauueYabrnlFh588EE+//xzZs+ezauvvspTTz2Vxq59icPGCkYSkyZNori4mLZt2zJ69Gh+/vOfc9ZZZ3HPPfewbNkyxo4dS4sWLaqHKrfccgsDBgyonp44cSIff/wxhx56KBdddBGTJk36Us9mxIgR3HnnndXXNm699VbuuusuevTowfXXX1+nC6z7i/weYF1t2LCBd999l5EjR9KkSRO++tWvcvnllzNjxowv1X3qqaeICPr371/dA5REaWkpCxYsKNRu7NKf5d/6tnQUFRUxe/bsGueNGTOGMWPG1Lrs9ddf/4Xpdu3a8fjjj+9ye7NmzfrC9HHHHcfixYvr2Nr6s27dOp5//nm+973v0bRpU5599lmmTp3K1KlTAdi+fTu5L67kboV/8sknNGnShNz35v5X+/btKSkpYdKkSYwePZotW7YwZcoUjj322C/U++STT7j22mv5/e9/D+R6gC+88AI/+MEPmDNnDn//93+fwV67Z2OWuV31AAG+/vWv07RpU1avXk2/fv1o2rQpK1asAL7cA3z00Ud56qmnKCoq4qijjuLggw9mwoQJX9jeLbfcwiWXXEJxcTEAV1xxBR988AFFRUUUFxdzzjnnZLPfVQl6ICktLY25c+fWOv+MC8Zl2Jr69cxDN+31sr1+OrZg7WgI5t8wtr6bcECQNC8iSncu92mUWQH8n+euru8mZObfTr9zr5bzaZSZZcJhY2aZcNiYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlonUwkbSIZJek/SmpEWSfpKUl0h6VVK5pIckNU7KmyTT5cn8Lnnrui4p/6Okfmm12czSk2bPZjtwWkQcB/QC+kvqDfwzMCEijgI+BIYm9YcCHyblE5J6SDoGuBDoAfQHJkra/SfMzGy/klrYRM6WZPLgZAjgNOA/k/IpwNnJ+KBkmmT+6cq95joImBYR2yNiOVAOnJRWu80sHales5HUSNJ8YB0wE1gKfBQRlUmVVUDHZLwjsBIgmb8R+Gp+eQ3LmFkDkWrYRMSOiOgFFJPrjXRPa1uShkmaK2nu+vXr09qMme2lTO5GRcRHwCzgL4E2kqreNi8GVifjq4FOAMn81sCf8strWCZ/G5MjojQiSouKilLZDzPbe2nejSqS1CYZbwp8F1hCLnTOTaoNAZ5IxsuSaZL5z0fuYztlwIXJ3aoSoCvwWlrtNrN0pPk9mw7AlOTO0VeAhyPi95IWA9Mk/RT4L+DXSf1fA/dLKgc2kLsDRUQskvQwsBioBEZExA7MrEFJLWwiYgFwfA3ly6jhblJEfAKcV8u6xgPjC91GM8uOnyA2s0w4bMwsEw4bM8uEw8bMMuGwMbNMOGzMLBMOGzPLhMPGzDLhsDGzTDhszCwTDhszy4TDxswy4bAxs0w4bMwsEw4bM8uEw8bMMuGwMbNMOGzMLBMOGzPLhMPGzDLhsDGzTDhszCwTDhszy4TDxswy4bAxs0w4bMwsEw4bM8tEamEjqZOkWZIWS1ok6eqkfKyk1ZLmJ8PAvGWuk1Qu6Y+S+uWV90/KyiVdm1abzSw9B6W47krgRxHxhqSWwDxJM5N5EyLi9vzKko4BLgR6AH8BPCupWzL7F8B3gVXA65LKImJxim03swJLLWwiYg2wJhnfLGkJ0HEXiwwCpkXEdmC5pHLgpGReeUQsA5A0LanrsDFrQDK5ZiOpC3A88GpSNFLSAkn3SmqblHUEVuYttiopq63czBqQ1MNGUgtgOjAqIjYBk4AjgV7kej7/WqDtDJM0V9Lc9evXF2KVZlZAqYaNpIPJBc2DEfEoQESsjYgdEfE58Cv+91RpNdApb/HipKy28i+IiMkRURoRpUVFRYXfGTPbJ2nejRLwa2BJRNyRV94hr9o5wFvJeBlwoaQmkkqArsBrwOtAV0klkhqTu4hclla7zSwdad6N+mvgMmChpPlJ2fXARZJ6AQFUAFcARMQiSQ+Tu/BbCYyIiB0AkkYCTwONgHsjYlGK7TazFKR5N+oPgGqYNWMXy4wHxtdQPmNXy5nZ/s9PEJtZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpaJOoWNpOfqUmZmVptdho2kQyS1A9pLaiupXTJ0ATruZtlOkmZJWixpkaSrk/J2kmZKeif52TYpl6S7JJVLWiDphLx1DUnqvyNpyL7utJllb3c9myuAeUD35GfV8ARw926WrQR+FBHHAL2BEZKOAa4FnouIrsBzyTTAAKBrMgwDJkEunIAxwMnAScCYqoAys4Zjl2ETEXdGRAkwOiK+FhElyXBcROwybCJiTUS8kYxvBpaQ6w0NAqYk1aYAZyfjg4D7IucVoI2kDkA/YGZEbIiID4GZQP+9210zqy8H1aVSRPybpL8CuuQvExH31WX55LTreOBV4LCIWJPMeh84LBnvCKzMW2xVUlZbuZk1IHUKG0n3A0cC84EdSXEAuw0bSS2A6cCoiNgkqXpeRISk2NNG17KdYeROvzjiiCMKsUozK6A6hQ1QChwTEXsUDJIOJhc0D0bEo0nxWkkdImJNcpq0LilfDXTKW7w4KVsNfGun8hd23lZETAYmA5SWlhYkwMyscOr6nM1bwOF7smLlujC/BpZExB15s8qAqjtKQ8hdbK4q/35yV6o3sDE53XoaOCO5G9YWOCMpM7MGpK49m/bAYkmvAdurCiPirF0s89fAZcBCSfOTsuuB24CHJQ0FVgDnJ/NmAAOBcmAbcHmyjQ2SbgZeT+qNi4gNdWy3me0n6ho2Y/d0xRHxB0C1zD69hvoBjKhlXfcC9+5pG8xs/1HXu1Gz026ImR3Y6no3ajO5u08AjYGDga0R0SqthpnZgaWuPZuWVePJhd9B5J4KNjOrkz1+6zt5wvdxck/2mpnVSV1PowbnTX6F3HM3n6TSIjM7INX1btSZeeOVQAW5Uykzszqp6zWby9NuiJkd2Or68axiSY9JWpcM0yUVp904Mztw1PUC8X+Qe53gL5Lhd0mZmVmd1DVsiiLiPyKiMhl+AxSl2C4zO8DUNWz+JOlSSY2S4VLgT2k2zMwOLHUNm78l98Lk+8Aa4FzgBym1ycwOQHW99T0OGJJ8lrPqu8C3kwshM7PdqmvP5tiqoIHcZx/IfebTzKxO6ho2X8n/iwZJz6auvSIzszoHxr8CL0t6JJk+DxifTpPM7EBU1yeI75M0FzgtKRocEYvTa5aZHWjqfCqUhIsDxsz2yh5/YsLMbG84bMwsEw4bM8uEw8bMMuGwMbNMOGzMLBMOGzPLhMPGzDLhsDGzTKQWNpLuTb5X/FZe2VhJqyXNT4aBefOuk1Qu6Y+S+uWV90/KyiVdm1Z7zSxdafZsfgP0r6F8QkT0SoYZAJKOAS4EeiTLTKz6KiDwC2AAcAxwUVLXzBqY1D4TEREvSupSx+qDgGkRsR1YLqkcOCmZVx4RywAkTUvq+h0tswamPq7ZjJS0IDnNqvpGTkdgZV6dVUlZbeVfImmYpLmS5q5fvz6NdpvZPsg6bCYBRwK9yH3L+F8LteKImBwRpRFRWlTkP/xgtr/J9Gt7EbG2alzSr4DfJ5OrgU55VYuTMnZRbmYNSKY9G0kd8ibPAaruVJUBF0pqIqkE6Aq8BrwOdJVUIqkxuYvIZVm22cwKI7WejaSpwLeA9pJWAWOAb0nqBQRQAVwBEBGLJD1M7sJvJTAiInYk6xkJPA00Au6NiEVptdnM0pPm3aiLaij+9S7qj6eG7xont8dnFLBpZlYP/ASxmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWCYeNmWXCYWNmmUgtbCTdK2mdpLfyytpJminpneRn26Rcku6SVC5pgaQT8pYZktR/R9KQtNprZulKs2fzG6D/TmXXAs9FRFfguWQaYADQNRmGAZMgF07AGOBk4CRgTFVAmVnDklrYRMSLwIadigcBU5LxKcDZeeX3Rc4rQBtJHYB+wMyI2BARHwIz+XKAmVkDkPU1m8MiYk0y/j5wWDLeEViZV29VUlZbuZk1MPV2gTgiAohCrU/SMElzJc1dv359oVZrZgWSddisTU6PSH6uS8pXA53y6hUnZbWVf0lETI6I0ogoLSoqKnjDzWzfZB02ZUDVHaUhwBN55d9P7kr1BjYmp1tPA2dIaptcGD4jKTOzBuagtFYsaSrwLaC9pFXk7irdBjwsaSiwAjg/qT4DGAiUA9uAywEiYoOkm4HXk3rjImLni85m1gCkFjYRcVEts06voW4AI2pZz73AvQVsmpnVAz9BbGaZcNiYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWiXoJG0kVkhZKmi9pblLWTtJMSe8kP9sm5ZJ0l6RySQsknVAfbTazfVOfPZtvR0SviChNpq8FnouIrsBzyTTAAKBrMgwDJmXeUjPbZ/vTadQgYEoyPgU4O6/8vsh5BWgjqUN9NNDM9l59hU0Az0iaJ2lYUnZYRKxJxt8HDkvGOwIr85ZdlZSZWQNyUD1tt09ErJZ0KDBT0n/nz4yIkBR7ssIktIYBHHHEEYVrqZkVRL30bCJidfJzHfAYcBKwtur0KPm5Lqm+GuiUt3hxUrbzOidHRGlElBYVFaXZfDPbC5mHjaTmklpWjQNnAG8BZcCQpNoQ4IlkvAz4fnJXqjewMe90y8waiPo4jToMeExS1fZ/GxFPSXodeFjSUGAFcH5SfwYwECgHtgGXZ99kM9tXmYdNRCwDjquh/E/A6TWUBzAig6aZWYr2p1vfZnYAc9iYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWCYeNmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpYJh42ZZcJhY2aZcNiYWSYcNmaWiQYTNpL6S/qjpHJJ19Z3e8xszzSIsJHUCPgFMAA4BrhI0jH12yoz2xMNImyAk4DyiFgWEZ8C04BB9dwmM9sDDSVsOgIr86ZXJWVm1kAoIuq7Dbsl6Vygf0T8XTJ9GXByRIzMqzMMGJZMfh34Y+YN3b32wAf13YgGwMepbvbX49Q5Iop2LjyoPlqyF1YDnfKmi5OyahExGZicZaP2lKS5EVFa3+3Y3/k41U1DO04N5TTqdaCrpBJJjYELgbJ6bpOZ7YEG0bOJiEpJI4GngUbAvRGxqJ6bZWZ7oEGEDUBEzABm1Hc79tF+fZq3H/FxqpsGdZwaxAViM2v4Gso1GzNr4Bw2ZpYJh02GJI3dafqrkmZJ2iLp7npqVsFI6iLpB3nTfSW9IakyeVaqtuVqrCepl6SXJS2StEDSBXnzSiS9mrwr91BylxJJYyWNTsYPkTRz5+O+P6rrsWvIx8RhkxJJjSU1T8b7SHoVuFLS65JOT6p9AtwIjK6vdhaKpKuAJ4GbJb0g6XDgXeAHwG93s3ht9bYB34+IHkB/4OeS2iTz/hmYEBFHAR8CQ3dqT2NgOjAvIsbu5W5lYg+PXYM9Jg3mblRDIelo4O+AwcnwX8AdwD8C3wYmAK0BImIr8AdJR9VPawtDUkvgJ+T+8R8LvABsjYj3k/mf72r5iKioqV5EvJ03/p6kdUCRpI3AacDFyewpwFhgUjJ9EPAQ8E5E7NdfCNjTY9eQj4l7NgUgqbmkyyX9AfgVsBg4NiL+K6nyKVAEEBEbI+LdempqWj4HAmgHufCIiM2F3ICkk4DGwFLgq8BHEVGZzN75Xbn/C3waEaMK2YaU7PWxa2jHxD2bwlgDLAD+LiL+u4b5w8h1cf9a0teB66p+mx8IImKrpB8CtwKHS+oJ3BQR2wqxfkkdgPuBIRHxuaTdLfIH4K8kdcvvCeyP9vbYNcRj4p5NYZxL7l2tRyXdJKlz/syIWBwRZ5Lr0r4E/Loe2piqiCgDzgP+hVwv7ke11ZU0XtJ8SfN3t15JrYD/B/xTRLySFP8JaCOp6pflzu/KvQiMAp5M/lPu1/bk2EHDPSYOmwKIiGci4gLgFGAj8ISkZyV1AUh+WwF8BrwBtKyPdqZFUou8gN0MLGEX+xgR/xQRvSKi127W2xh4DLgvIv4zb/kAZpELeYAhwBM7bWM6cDvwVN4F1P3Onh67Bn1MIsJDCgO5D351SsYnAv+f3Dd55gPfyqtXAWwAtpA7zz6mvtu+F/vaFniK3Auzy4HZ5K4XfDPZp63kfvMuqmX5GusBl5IL6Pl5Q69k3teA14By4BGgSVI+Fhidt+6x5H6rH1Lfx6kQx64hHxO/rpAhSWNjP78Nuy+Snty3IuI39duShufP4dj5NCpbL9R3A1L2EbnftLbnDvhj556NmWXCPRszy4TDxswy4bCxgpC0I3l25s3kBcK/qsMyW7Jom+0f/ASxFcrHkTw3I6kfuSdiT63fJtn+xD0bS0Mrcm8dVz209lzS21ko6Ut/XLC2OslnF5ZI+lXySYVnJDVN5h2VPDhZ1ZM6Min/x+TN+gWSfpLhPttu+G6UFYSkHcBC4BCgA3BaRMxLHp9vFhGbJLUHXgG6RkRI2hIRLWqrA3Qm94BaaUTMl/QwUBYRDySf7LgtIh6TdAi5X5x9yD1BewUgcn+B418i4sUsj4XVzKdRVij5p1F/CdyXvKYh4BZJfcm94dwROAx4P2/Z2uoALI+IqudP5gFdks8ydIyIxwAi4pNku2cAZ5D7rAdAC3Kh5bDZDzhsrOAi4uWkh1IEDEx+nhgRn0mqINf7yXfJLupsz6u3A2i6i00LuDUifrnve2GF5ms2VnCSupP7+15/IvehsHVJiHyb3KnRzupSp1rkvveyStLZyfaaSGpG7u+K/a2kFkl5R0mHFmzHbH1YiEoAAABxSURBVJ+4Z2OF0jTvkxEi952VHZIeBH4naSEwF6jpez91qbOzy4BfShpH7sXE8yLimeRLiS8n33fZQu7FxXX7smNWGL5AbGaZ8GmUmWXCYWNmmXDYmFkmHDZmlgmHjZllwmFjZplw2JhZJhw2ZpaJ/wF+FPm/Xjli+QAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 288x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#provide boxplots of categorical variables\n",
    "for col in df.columns[df.dtypes == 'category']:\n",
    "    labeled_barplot(df, col, perc=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "yNsneIZDc2LK",
    "outputId": "03114d17-8192-40b0-9320-addbf7080d6a"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-9b6f92cd-24b6-46b0-81a2-a1dd89bb5140\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>unique</th>\n",
       "      <th>top</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Geography</th>\n",
       "      <td>10000</td>\n",
       "      <td>3</td>\n",
       "      <td>France</td>\n",
       "      <td>5014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gender</th>\n",
       "      <td>10000</td>\n",
       "      <td>2</td>\n",
       "      <td>Male</td>\n",
       "      <td>5457</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Balance</th>\n",
       "      <td>10000</td>\n",
       "      <td>3</td>\n",
       "      <td>&lt;$1</td>\n",
       "      <td>3617</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HasCrCard</th>\n",
       "      <td>10000</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>7055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IsActiveMember</th>\n",
       "      <td>10000</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>5151</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9b6f92cd-24b6-46b0-81a2-a1dd89bb5140')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-9b6f92cd-24b6-46b0-81a2-a1dd89bb5140 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-9b6f92cd-24b6-46b0-81a2-a1dd89bb5140');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "                count unique     top  freq\n",
       "Geography       10000      3  France  5014\n",
       "Gender          10000      2    Male  5457\n",
       "Balance         10000      3     <$1  3617\n",
       "HasCrCard       10000      2       1  7055\n",
       "IsActiveMember  10000      2       1  5151"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#provide descriptive statistics for categorical variables\n",
    "df.describe(include = ['category']).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "8zWkhPOAdBn6",
    "outputId": "a7e3842c-937f-4888-d034-02b325fc6d3a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "France     5014\n",
      "Germany    2509\n",
      "Spain      2477\n",
      "Name: Geography, dtype: int64\n",
      " \n",
      "Male      5457\n",
      "Female    4543\n",
      "Name: Gender, dtype: int64\n",
      " \n",
      "<$1        3617\n",
      "$1-120K    3202\n",
      "$120K      3181\n",
      "Name: Balance, dtype: int64\n",
      " \n",
      "1    7055\n",
      "0    2945\n",
      "Name: HasCrCard, dtype: int64\n",
      " \n",
      "1    5151\n",
      "0    4849\n",
      "Name: IsActiveMember, dtype: int64\n",
      " \n"
     ]
    }
   ],
   "source": [
    "#provide counts for each unique variable in each categorical variable\n",
    "for colname in df.dtypes[df.dtypes == 'category'].index:\n",
    "    print(df[colname].value_counts(dropna=False))\n",
    "    print(\" \")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "wsxOJ9cGdPpq"
   },
   "source": [
    "#### Observations\n",
    "- Of the sample population of 10,000 bank account owners:\n",
    "  - %55 are male\n",
    "  - ~70% have a credit card\n",
    "  - ~52% are active members (i.e., use the bank regularly)\n",
    "  - 50% live in France, 25% live in Germany, and 25% live in Spain; so, the population is biased towards French people.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "7upAvpE7doW-"
   },
   "source": [
    "## Bivariate analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 432
    },
    "id": "uIE2C_AHdlt_",
    "outputId": "a44dda63-731f-4864-9252-adfb3ca3aef8"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1008x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create correlation heat map for numerical variables\n",
    "numeric_columns = df.select_dtypes(include=np.number).columns.tolist()\n",
    "\n",
    "#correlation heatmap\n",
    "plt.figure(figsize=(14, 7))\n",
    "sns.heatmap(\n",
    "    df[numeric_columns].corr(),\n",
    "    annot=True,\n",
    "    vmin=-1,\n",
    "    vmax=1,\n",
    "    fmt=\".2f\",\n",
    "    cmap='viridis'\n",
    ")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "cfxKLjKWFvkP",
    "outputId": "3871e9fb-221f-40b1-f3af-a8b911f439b3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Geography  Exited\n",
      "France     0        83.85\n",
      "           1        16.15\n",
      "Germany    0        67.56\n",
      "           1        32.44\n",
      "Spain      0        83.33\n",
      "           1        16.67\n",
      "Name: Exited, dtype: float64\n",
      " \n",
      "Gender  Exited\n",
      "Female  0        74.93\n",
      "        1        25.07\n",
      "Male    0        83.54\n",
      "        1        16.46\n",
      "Name: Exited, dtype: float64\n",
      " \n",
      "Balance  Exited\n",
      "<$1      0        86.18\n",
      "         1        13.82\n",
      "$1-120K  0        75.89\n",
      "         1        24.11\n",
      "$120K    0        75.95\n",
      "         1        24.05\n",
      "Name: Exited, dtype: float64\n",
      " \n",
      "HasCrCard  Exited\n",
      "0          0        79.19\n",
      "           1        20.81\n",
      "1          0        79.82\n",
      "           1        20.18\n",
      "Name: Exited, dtype: float64\n",
      " \n",
      "IsActiveMember  Exited\n",
      "0               0        73.15\n",
      "                1        26.85\n",
      "1               0        85.73\n",
      "                1        14.27\n",
      "Name: Exited, dtype: float64\n",
      " \n"
     ]
    }
   ],
   "source": [
    "#provide ratios of the dependent variable for all categorical variables\n",
    "for col in df.columns[df.dtypes == 'category']:\n",
    "    print(df.groupby(col)[\"Exited\"].value_counts(normalize=True).mul(100))\n",
    "    print(\" \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "WmV5Qg1NKYcp",
    "outputId": "282d7f3c-45f2-4acd-b95d-b2e6e1bd5030"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Tenure  Exited\n",
       "0       0        77.00\n",
       "        1        23.00\n",
       "1       0        77.58\n",
       "        1        22.42\n",
       "2       0        80.82\n",
       "        1        19.18\n",
       "3       0        78.89\n",
       "        1        21.11\n",
       "4       0        79.47\n",
       "        1        20.53\n",
       "5       0        79.35\n",
       "        1        20.65\n",
       "6       0        79.73\n",
       "        1        20.27\n",
       "7       0        82.78\n",
       "        1        17.22\n",
       "8       0        80.78\n",
       "        1        19.22\n",
       "9       0        78.35\n",
       "        1        21.65\n",
       "10      0        79.39\n",
       "        1        20.61\n",
       "Name: Exited, dtype: float64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#provide ratios of the dependent variable for all distinct values of Tenure\n",
    "df.groupby('Tenure')[\"Exited\"].value_counts(normalize=True).mul(100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Y6O4oH0-d4Z4"
   },
   "source": [
    "#### Observations\n",
    "- The only variables which appear correlated are *Age* and *Exited*.\n",
    "  - The low positive correlation signifies that, as a person ages, they are slightly more likely to leave the bank.\n",
    "- A higher ratio of females in the population exited the bank than males.\n",
    "- A higher ratio of people with a balance less than $1k in the population exited the bank than those with a balance greater than that amount.\n",
    "- A lower ratio of people who were categorized as \"Active Members\" exited the bank than those who were not.\n",
    "- Those with less than one year of tenure with the bank have the highest ratio of exits and those with only one year of tenure have the second highest ratio of exits."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TJl7BYubYB_T"
   },
   "source": [
    "## Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "id": "nc4l02acWR0a"
   },
   "outputs": [],
   "source": [
    "#create dummy variables for the categorical features\n",
    "df_train = pd.get_dummies(df,drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "USodcIc-IKuS",
    "outputId": "928b2fdf-14b9-4e68-bbbd-07e80c19736f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 10000 entries, 0 to 9999\n",
      "Data columns (total 13 columns):\n",
      " #   Column             Non-Null Count  Dtype  \n",
      "---  ------             --------------  -----  \n",
      " 0   CreditScore        10000 non-null  int32  \n",
      " 1   Age                10000 non-null  int32  \n",
      " 2   Tenure             10000 non-null  int32  \n",
      " 3   NumOfProducts      10000 non-null  int32  \n",
      " 4   EstimatedSalary    10000 non-null  float32\n",
      " 5   Exited             10000 non-null  int32  \n",
      " 6   Geography_Germany  10000 non-null  uint8  \n",
      " 7   Geography_Spain    10000 non-null  uint8  \n",
      " 8   Gender_Male        10000 non-null  uint8  \n",
      " 9   Balance_$1-120K    10000 non-null  uint8  \n",
      " 10  Balance_$120K      10000 non-null  uint8  \n",
      " 11  HasCrCard_1        10000 non-null  uint8  \n",
      " 12  IsActiveMember_1   10000 non-null  uint8  \n",
      "dtypes: float32(1), int32(5), uint8(7)\n",
      "memory usage: 302.9 KB\n"
     ]
    }
   ],
   "source": [
    "#review training data\n",
    "df_train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "id": "-NV7_VuZdAHr"
   },
   "outputs": [],
   "source": [
    "#seperate features from labels\n",
    "X = df_train.drop('Exited',axis=1)   # Predictor feature columns\n",
    "Y = df_train['Exited']   # Predicted class (1=Exited, 0=Not Exited)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "id": "xhihQaPY6LfN"
   },
   "outputs": [],
   "source": [
    "#split the data into training and test sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.20, random_state=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8-PHB5E46w-0"
   },
   "source": [
    "- NOTE: Do not scale the dependent variable in a classification problem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "id": "kohyyWDkf75J"
   },
   "outputs": [],
   "source": [
    "#define scaler\n",
    "scaler = preprocessing.StandardScaler()\n",
    "#fit the scaler on the independent variables for both train and test\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "id": "ium-ktErgOLt"
   },
   "outputs": [],
   "source": [
    "#convert the data elements into tensors, as we need tensors for different tensorflow based operations\n",
    "x_train1=tf.convert_to_tensor(X_train)\n",
    "y_train1=tf.convert_to_tensor(y_train)\n",
    "x_test1=tf.convert_to_tensor(X_test)\n",
    "y_test1=tf.convert_to_tensor(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "GOUVOqYeg2mj",
    "outputId": "b286df82-bd8c-49eb-d312-b8a11f04322a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8000, 12)\n",
      "(8000,)\n",
      "(2000, 12)\n",
      "(2000,)\n"
     ]
    }
   ],
   "source": [
    "#print the shape of training tensors\n",
    "print(x_train1.shape)\n",
    "print(y_train1.shape)\n",
    "\n",
    "#print the shape of training tensors\n",
    "print(x_test1.shape)\n",
    "print(y_test1.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Z03WqvqcN-eq"
   },
   "source": [
    "## Metric selection\n",
    "AUC should be the target metric for the following reasons:\n",
    "  - AUC provides an aggregate measure of performance across all possible classification thresholds.\n",
    "  - AUC is scale-invariant (i.e., measures how well predictions are ranked, rather than their absolute values).\n",
    "  - AUC is classification-threshold-invariant (i.e., measures the quality of the model’s predictions irrespective of what classification threshold is chosen)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "id": "eczyX-SVO6ww"
   },
   "outputs": [],
   "source": [
    "#set the metric for training the ANNs\n",
    "metric = tf.keras.metrics.AUC()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "qliScvpJ_3gU"
   },
   "source": [
    "## Models 1 & 2: Baseline models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ki8eXV1oFrS6"
   },
   "source": [
    "### Building"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "id": "d9ee78e9"
   },
   "outputs": [],
   "source": [
    "backend.clear_session()\n",
    "#set the seed for random number generators\n",
    "np.random.seed(42)\n",
    "import random\n",
    "random.seed(42)\n",
    "tf.random.set_seed(42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "id": "lAW8tWDIWEDp"
   },
   "outputs": [],
   "source": [
    "#set the layers within the neural network\n",
    "model_1 = Sequential()\n",
    "model_1.add(Dense(9, activation='relu',input_shape=(x_train1.shape[1],))) #input layer\n",
    "model_1.add(Dense(6, activation='relu')) #hidden layer\n",
    "model_1.add(Dense(1, activation='sigmoid')) #output layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "id": "Tw6oZMXRZKwg"
   },
   "outputs": [],
   "source": [
    "#set the optimizer, loss function, and metric\n",
    "adam = Adam(learning_rate=1e-3)\n",
    "model_1.compile(loss='binary_crossentropy', optimizer=adam, metrics=metric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "mQ372NV2RKai",
    "outputId": "d4fcd8ae-cd4d-4848-eeac-b9f62bc3bdf0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " dense (Dense)               (None, 9)                 117       \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 6)                 60        \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 1)                 7         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 184\n",
      "Trainable params: 184\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#show a summary of the ANN\n",
    "model_1.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "l3Y9dNV0SzDM"
   },
   "source": [
    "### Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "id": "8FIMSUF3aDpu"
   },
   "outputs": [],
   "source": [
    "#train model across select number of epochs\n",
    "epochs=30\n",
    "history = model_1.fit(x_train1, y_train1, validation_split=0.2, epochs=epochs, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 350
    },
    "id": "rC2n_0_wmcJH",
    "outputId": "14b7b184-6b41-43a2-d2cc-2d5d41e269f9"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot progression of losses across training epochs\n",
    "N = epochs\n",
    "plt.figure(figsize=(7,5))\n",
    "plt.plot(np.arange(0, N), history.history[\"loss\"], label=\"training\")\n",
    "plt.plot(np.arange(0, N), history.history[\"val_loss\"], label=\"validation\")\n",
    "\n",
    "plt.title(\"Training and validation losses across epochs\")\n",
    "plt.xlabel(\"Epoch\")\n",
    "plt.ylabel(\"Losses\")\n",
    "plt.legend(loc=\"best\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 294
    },
    "id": "JHGaN3dfnMLt",
    "outputId": "de95e30a-7238-44fe-e88e-7ecbae966c7a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot progression of accuracy across training epochs\n",
    "loss, accuracy  = model_1.evaluate(x_train1, y_train1, verbose=False)\n",
    "    \n",
    "plt.plot(history.history['auc'])\n",
    "plt.plot(history.history['val_auc'])\n",
    "plt.title('Model AUC across epochs')\n",
    "plt.ylabel('AUC')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['training', 'validation'], loc='best')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Yh6ZElYvsuMD"
   },
   "source": [
    "- The ANN trained appears not to overfit the training data and to have reached a steady state of performance prior to the final epoch."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hbywzKpm-cze"
   },
   "source": [
    "### Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 350
    },
    "id": "g2w-1om3oONH",
    "outputId": "f211d201-c803-441a-ba9d-bb966dcbc7dc"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot ROC curve on testing data\n",
    "roc_auc_test_1 = roc_auc_score(y_test1, model_1.predict(x_test1))\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_1.predict(x_test1))\n",
    "plt.figure(figsize=(7, 5))\n",
    "plt.plot(fpr, tpr, label=\"Model_1 (area = %0.2f)\" % roc_auc_test_1)\n",
    "plt.plot([0, 1], [0, 1], \"r--\")\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.01])\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate\")\n",
    "plt.title(\"Receiver operating characteristic\")\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "TWSLymJetItB",
    "outputId": "1eb71160-eb88-4e02-f79f-5f37d1e0cb3d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.88      0.96      0.92      1585\n",
      "           1       0.75      0.50      0.60       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.81      0.73      0.76      2000\n",
      "weighted avg       0.85      0.86      0.85      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set baseline threshold\n",
    "threshold = 0.5\n",
    "#predict the labels of training data\n",
    "y_pred_test1=np.round(model_1.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(y_test1,y_pred_test1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "AtZyN3Usx6jw",
    "outputId": "4bc025d5-5138-483b-9ed2-9f859f8f8fbc"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test1)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ztzIEhN5AWvw"
   },
   "source": [
    "### Optimal threshold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2nMho3a2qK-F",
    "outputId": "2a3424c0-513e-4edb-b06a-87826a195413"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.18896937\n"
     ]
    }
   ],
   "source": [
    "#optimal threshold as per AUC-ROC curve\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_1.predict(x_test1))\n",
    "\n",
    "optimal_idx = np.argmax(tpr - fpr)\n",
    "optimal_threshold_auc_roc = thresholds[optimal_idx]\n",
    "print(optimal_threshold_auc_roc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "78fe9gkN2EaS",
    "outputId": "7869e731-9774-464f-f7c9-992298d1ef30"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.77      0.84      1585\n",
      "           1       0.47      0.79      0.59       415\n",
      "\n",
      "    accuracy                           0.78      2000\n",
      "   macro avg       0.70      0.78      0.72      2000\n",
      "weighted avg       0.84      0.78      0.79      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set optimal threshold\n",
    "threshold = optimal_threshold_auc_roc\n",
    "#predict the labels of training data\n",
    "y_pred_test2=np.round(model_1.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "print(classification_report(y_test1,y_pred_test2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "JeXdQreX4TzU",
    "outputId": "169bdd87-927b-4ae6-8143-552f0fc7a7c6"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAacAAAE9CAYAAACvEq29AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAbpklEQVR4nO3de7hVVb3/8fd3bwQvGCgmKmpSoomlaWYcPZlpGWilv7LSrNQojh21vFRinXM81ak0TY/1U4vU1PJaamJqauQ1RbzmNZXwBmKYF8wQgb2/54810SXCZrNkseZc+/3ymY9rjTnXnGP5+PDhO+ZYY0ZmIklSmXS0ugOSJC3KcJIklY7hJEkqHcNJklQ6hpMkqXQMJ0lS6fRrdQeWpPupTZzjrhXm/V8e1+ouqI+54eKvxfI6V6N/Xnas89By68PyZuUkSVqsiDg9ImZFxL11bcdGxF8i4u6IuDgiBtftOzIipkbEgxHx4br20UXb1IgY35trG06SVHHdDf7TC2cAoxdpuxp4R2ZuATwEHAkQESOBvYDNi8+cHBGdEdEJnASMAUYCexfH9qi0w3qSpN7pyl4FzessLQAy8/qI2GiRtqvq3k4G9ixe7w6cl5kvA49ExFRg22Lf1MycBhAR5xXH3t/Tta2cJKniusmGtuXgC8AVxethwBN1+6YXbUtq75GVkyRVXC+H6F4nIsYB9bOBJmTmhF5+9lvAAuDshi6+FIaTJFVcV4MLeBdB1KswqhcR+wEfAXbOV1cPnwFsUHfY+kUbPbQvkcN6klRxK3JYLyJGA98APpaZc+p2TQT2iogBETEcGAFMAW4FRkTE8IjoT23SxMSlXcfKSZIqrmv53D96nYg4F9gRWCsipgNHUZudNwC4OiIAJmfmAZl5X0RcQG2iwwLgwMzsKs5zEHAl0Amcnpn3Le3ahpMkVdxymtzwOpm592KaT+vh+O8B31tM++XA5ctybcNJkiqu0XtOZWY4SVLFNTZXr9wMJ0mquGbdc2olw0mSKq6r/bLJcJKkqnNYT5JUOl2U9skXDTOcJKniuh3WkySVTTtWTi5fJEkqHSsnSaq4dqycDCdJqrjuNJwkSSVj5SRJKp2uNpw+YDhJUsU5rCdJKh2H9SRJpdOVDutJkkqm23tOkqSycVhPklQ6DutJkkqn28pJklQ2/s5JklQ6DutJkkrH2XqSpNLpcoUISVLZtOM9p/b7RpKkyrNykqSK63ZChCSpbNpxWM9wkqSKc0KEJKl0nEouSSodf4QrSSod19aTJJWOlZMkqXScrSdJKp1uZ+tJksrGykmSVDquECFJKp0uZ+tJksqmHSun9vtGktTHdBENbUsTEadHxKyIuLeubc2IuDoiHi7+vUbRHhHx44iYGhF3R8TWdZ/Ztzj+4YjYtzffyXCSpIrrzo6Gtl44Axi9SNt4YFJmjgAmFe8BxgAjim0ccArUwgw4CngvsC1w1MJA64nhJEkV15UdDW1Lk5nXA88u0rw7cGbx+kxgj7r2s7JmMjA4ItYFPgxcnZnPZuZzwNW8PvBex3CSJC2LoZk5s3j9FDC0eD0MeKLuuOlF25Lae2Q4SVLFdRMNbRExLiJuq9vGLct1MzOBbMZ3craeJFVco2vrZeYEYMIyfuxvEbFuZs4shu1mFe0zgA3qjlu/aJsB7LhI+7VLu4iVkyRVXHdGQ1uDJgILZ9ztC1xS1/75YtbeKGB2Mfx3JbBLRKxRTITYpWjrkZWTJFVcs5YviohzqVU9a0XEdGqz7o4GLoiIscBjwKeKwy8HdgWmAnOA/QEy89mI+C5wa3HcdzJz0UkWr2M4SVLFNWvh18zcewm7dl7MsQkcuITznA6cvizXNpwkqeJ8TLskqXS6fGSGJKlsfJ6TJKl02nHhV8NJkirOR2aoJb51NFx7M6y5Blx6Rq3t2FPgmptgpX6wwXrw/fHwptXhT7fC8RNg/nxYaSX4+pdhVLE28P/+HC65El54EW7/fcu+jiqk/0qd/OR7e9G/XyednR1ce/NDnH7eTQB8aZ9/5QPbbUJXd/Lb39/FhZfd+crn3r7xOpxy9Gf49o9+x7U3P9Sq7vcZDuupJfYYA5/5OIz//qtt220Dh34J+vWD434KE86Grx0AawyCU34Aa68FD02DL30drruw9pkdt6udZ8w+rfkeqp5587s45L8u4KW58+ns7ODk7+/N5DseYaP1h7D2kNXZ56DTyYTBg1Z95TMdHcEBn9+BW+96tHUd72PacViv/b5RG3rPljB49de2bf+eWjABbDkS/vZ07fXITWrBBDBiOLz8MsybV3v/rs1h7SErps9qHy/NnQ9Av84O+nV2QCa7j96SMy64mSxWVXt+9pxXjv/Erltx3c0PvaZNzdXo2npl1rTKKSLeTm0J9YWrz84AJmbmA826Zl910eUwZqfXt191HWy2CfTvv+L7pPbR0RGcetznGLbOYC6+4i7uf/gphq0zmJ3+dVN2eO8Inn/hJU48dRLTZz7PWmsOZIdRI/jKf57PZgct9akIWk7acSp5UyqniDgCOA8IYEqxBXBuRIzv6bNaNj/9JXR2wkc/9Nr2hx+BH/0Mvn14a/ql9tHdnXzhsLP4xBd/xmYj1mH4hmuxUr9O5s3r4ktf/xWXXn0344sg+srYD3DKWde/UlFpxWjiwwZbplmV01hg88ycX98YEccD91Fbm+l1iuXaxwGc8sO1Gfe5QU3qXnu4+Aq49ib4xQkQdX9xemoWHPwfcPQ3YcOlPjVF6p0X57zMnfc+wXu32oinn/kH109+GIDrJz/MkUU4bfq2dfjvwz8CwKDVV2HUu99KV1c3N0yZ2rJ+9wVOiOi9bmA9aosC1lu32LdY9cu3dz+1iX/36sENt8Bp58JZP4ZVVn61/YV/wAHj4bB/g63f2br+qT0MftMqLFjQzYtzXqZ//35ss+VbOOfiKdwwZSpbvXMDZk6azbs234AnnnwOgE8f8PNXPvvNg0dz023TDCY1pFnhdAgwKSIe5tUnIG4IbAwc1KRrtq3Dvw1T7oLnZ8OOe8JB+8PPz65NdBhbDNttORL++3A4+2J4fAaccmZtAzj1OBiyRm36+WWT4KW5tfPsuVvtXNKSDFljNb75lTF0dnQQHcE1f3qQm26bxt33z+C/Dt2NT3303bw0dz7HnLzUJyCoico+uaERkU0aHI6IDmBbXjsh4tbM7OrN562ctCK9/8vL9ABQ6Q274eKvLbdE2XvyuIb+vDx31ITSplrTZutlZjcwuVnnlyTVlH1yQyP8Ea4kVZwTIiRJpdOO95wMJ0mqOCsnSVLpGE6SpNIxnCRJpWM4SZJKxwkRkqTSsXKSJJWO4SRJKh3DSZJUOoaTJKl00nCSJJWNs/UkSaXTjsN67bfOuiSp8qycJKnivOckSSqddhzWM5wkqeKsnCRJpWPlJEkqncxW92D5M5wkqeL8nZMkqXS85yRJKh3vOUmSSqcd7zm5QoQkVVxmNLQtTUQcGhH3RcS9EXFuRKwcEcMj4paImBoR50dE/+LYAcX7qcX+jd7IdzKcJKnimhFOETEM+AqwTWa+A+gE9gKOAU7IzI2B54CxxUfGAs8V7ScUxzXMcJKkiuvOaGjrhX7AKhHRD1gVmAnsBPym2H8msEfxevfiPcX+nSOi4ZthhpMkVVxmY1vP58wZwHHA49RCaTZwO/B8Zi4oDpsODCteDwOeKD67oDh+SKPfyXCSpIprdFgvIsZFxG1127iF54yINahVQ8OB9YDVgNEr6js5W0+SKq7R3zll5gRgwhJ2fxB4JDOfBoiIi4DtgcER0a+ojtYHZhTHzwA2AKYXw4CDgGca6hhWTpJUedngthSPA6MiYtXi3tHOwP3ANcCexTH7ApcUrycW7yn2/zGz8UnuhpMk6XUy8xZqExvuAO6hlhcTgCOAwyJiKrV7SqcVHzkNGFK0HwaMfyPXd1hPkiquWcsXZeZRwFGLNE8Dtl3MsXOBTy6vaxtOklR1bbhChOEkSRXnwq+SpNJpx7X1DCdJqjgrJ0lS+RhOkqSycVhPklQ+hpMkqWy85yRJKh8rJ0lS2Vg5SZLKx8pJklQ+Vk6SpLKxcpIklY7hJEkqnTacEOHDBiVJpWPlJEkV5/JFkqTyMZwkSaXThvecehVOEbEdsFH98Zl5VpP6JElaBtEXK6eI+CXwNuAuoKtoTsBwkqQy6IvhBGwDjMxsx1tuktQG2nBYrzdTye8F1ml2RyRJDcoGtxJbYuUUEZdS6/7qwP0RMQV4eeH+zPxY87snSVqqkgdNI3oa1jtuhfVCktS4vhROmXkdQEQck5lH1O+LiGOA65rcN0lSb/TRe04fWkzbmOXdEUlSYyIb28qsp3tOXwb+HXhbRNxdt2t14KZmd0yS1EslD5pG9HTP6RzgCuAHwPi69n9k5rNN7ZUkqU/r6Z7TbGB2RByxyK6BETEwMx9vZsc+vN6WzTy99Bqdu3Qt/SCppMo+RNeI3vwI9zJqRWMAKwPDgQeBzZvYL0lSb7XhhIilhlNmvrP+fURsTe1elCSpDNqwclrmhw1m5h3Ae5vQF0mSgN4t/HpY3dsOYGvgyab1SJK0bNqwcurNPafV614voHYP6sLmdEeStKz63ISIiOgEVs/Mr62g/kiSllVfCqeI6JeZCyJi+xXZIUnSMupL4QRMoXZ/6a6ImAj8Gvjnwp2ZeVGT+yZJ6oU+N6xXWBl4BtiJV3/vlIDhJEll0Md+57R2MVPvXl4NpYXaMKclqaKa+CdyRAwGTgXeUVzpC9QWYjgf2Ah4FPhUZj4XEQGcCOwKzAH2K35+tMx6+p1TJzCw2Fave71wkySVQJNXJT8R+H1mvh3YEniA2nqrkzJzBDCJV9dfHQOMKLZxwCmNfqeeKqeZmfmdRk8sSVpBmlQ5RcQgYAdgP4DMnAfMi4jdgR2Lw84ErgWOAHYHzsrMBCZHxOCIWDczZy7rtXuqnNpvEFOS2lATK6fhwNPALyLizog4NSJWA4bWBc5TwNDi9TDgibrPTy/alllP4bRzIyeUJK1g2dgWEeMi4ra6bdwiZ+5Hbdb2KZm5FbUZ2/WPUKKokpZ77dbTIzN8ZpMkVUGD0ZCZE4AJPRwyHZiembcU739DLZz+tnC4LiLWBWYV+2cAG9R9fv2ibZkt88KvkqRyadawXmY+BTwREZsWTTsD9wMTgX2Ltn2BS4rXE4HPR80oYHYj95ugd79zkiT1XQcDZ0dEf2AasD+1wuaCiBgLPAZ8qjj2cmrTyKdSm0q+f6MXNZwkqeqa+DunzLwL2GYxu143L6G4/3Tg8riu4SRJFdeOyxd5z0mSVDpWTpJUdW1YORlOklR1hpMkqWza8Z6T4SRJVWc4SZLKxspJklQ+hpMkqXQMJ0lS2TisJ0kqH8NJklQ6hpMkqWwc1pMklY/hJEkqGysnSVL5GE6SpNIxnCRJZROt7kATGE6SVHVtWDn5JFxJUulYOUlSxTlbT5JUPoaTJKl0DCdJUtk4rCdJKh/DSZJUNlZOkqTyMZwkSWVj5SRJKh/DSZJUOoaTJKlsHNaTJJWP4SRJKpvI9ksnw0mSqq79sslwkqSq856TJKl82jCcfNigJKl0rJwkqeIc1pMklU8bhpPDepJUcZGNbb06d0RnRNwZEb8r3g+PiFsiYmpEnB8R/Yv2AcX7qcX+jd7IdzKcJKnqssGtd74KPFD3/hjghMzcGHgOGFu0jwWeK9pPKI5rmOEkSRXXrMopItYHdgNOLd4HsBPwm+KQM4E9ite7F+8p9u9cHN8Qw0mSqi6zsW3p/hf4BtBdvB8CPJ+ZC4r304FhxethwBO17uQCYHZxfEMMJ0mquEYrp4gYFxG31W3jXjlnxEeAWZl5eyu+k7P1JKnqGpytl5kTgAlL2L098LGI2BVYGXgTcCIwOCL6FdXR+sCM4vgZwAbA9IjoBwwCnmmsZ4ZT5X38kN0YM3ZnMpNH73mcY79wMvNfns/+/7M3O+w5iu6ubi796VX89idXtLqrqqD+K3Vy4nGfYaWV+tHZ2cF1NzzIGb+6kW994yNsusk6dC3o5oEHZ/KjH19JV1c3AwcO4IhDd2W99QYzb94Cfnj8FTzy2N9b/TXaXnQv/ZhllZlHAkcCRMSOwNcyc5+I+DWwJ3AesC9wSfGRicX7m4v9f8xsfEVaw6nChqy3JnscvCtf3PxQ5s2dx3+cdygf2Gt7IuDN6w/hC5sdQmYy+M1vanVXVVHz5ndx2BHn8dLc+XR2dvCTH+3DlNum8Ydr7ud7P/wdAP85/qPsNnoLJl52F5/d61+YOm0W//ndi9lw/TX56oEf4vAjz2/xt+gDVuzvnI4AzouI/wHuBE4r2k8DfhkRU4Fngb3eyEUMp4rr7NfBgFX6s2D+AgasOoBnnnyW/b67Nz/Y50QW/qXl+adfaHEvVWUvzZ0PQL9+HfTr10Fmcsut017Z/8CDM3nzWqsD8JYN1+Kc8ycD8Pj0Z1ln6CDWGLwqzz0/Z8V3vA9p9goRmXktcG3xehqw7WKOmQt8cnld0wkRFfbMk8/ymx9dytmPncL5T/6cf86ew+1X3816bxvKjp/ejpOmHM33LvsmwzZep9VdVYV1dASnnrQfvz3vYG6741EeeHDmK/s6OzvYZefNmXLbIwD8ddosdth+EwDevsm6rDN00CvBpSZq3my9llnh4RQR+6/oa7argYNX418+9h4+99YD2WvYOFZebQA77/M+VhqwEvPmzuPAbcdzxal/4PDT/r3VXVWFdXcnXzzwDD752ZPZbNN1Gf6WtV7Zd+hBu3D3PdO5577pAJxzwWQGDlyZU0/aj4/vvjUP//VvdHeX+w/BdtDMFSJapRWV07eXtKN+WuP0nLakw1TY+oPv5KlHZzH77y/QtaCLGy++hZHbbcrT05/hxoumAHDjxVN46xZvaXFP1Q5e/OfL3Pnnx9l2m7cCsO8+2zN40CqcNGHSK8fMmTOPY46/nC8eeAbfP/YyBg9alSefer5VXe47mrtCREs05Z5TRNy9pF3A0CV9rn5a44c6Plny/3StN+vxv7PZe0cwYJX+vPzSPLba6Z08dPtfmfPCS2z5gc156hez2OL9I5n+0JOt7qoqatCgVeha0M2L/3yZ/v37sc3WG3HuBbew2+gteM+7h3PY+PNeMzo0cLUBzH15PgsWdLPb6C358z1PMGfOvNZ9gT6i7FVQI5o1IWIo8GFq6y7VC+CmJl2zz/nLlKnccOFkTr79h3Qt6OKvdz7K5RP+QP9V+nPkr77KJw75CC+9OJfjv/TTVndVFTVkzYEcefhudHQGHRFcc/1fuHnKX5l02dd56m+zOfmEzwJw/Z8e4qxzbmLDDYdw5OG7kSSPPvZ3fniCP2FYIUp+/6gR8QamoS/5pBGnAb/IzBsXs++czPzM0s5h5aQVaf4u27S6C+pjrv39EQ2vO7eo9+1xbEN/Xt7w268vtz4sb02pnDJzbA/7lhpMkqTec1hPklQ+hpMkqWysnCRJ5dOGvyUznCSp6tovmwwnSao6h/UkSeXThr9zMpwkqeKsnCRJ5WM4SZLKJhzWkySVThMe095qhpMkVZyVkySpfNovmwwnSao8KydJUtk4lVySVD5tWDl1tLoDkiQtyspJkiounEouSSqdNhzWM5wkqeraL5sMJ0mqOn+EK0kqH8NJklQ6ToiQJJWNw3qSpPIxnCRJpWM4SZJKx3tOkqSy8Z6TJKl8DCdJUukYTpKk0jGcJEml44QISVLZtOOECB82KElarIjYICKuiYj7I+K+iPhq0b5mRFwdEQ8X/16jaI+I+HFETI2IuyNi60avbThJUtVlNrYt3QLg8MwcCYwCDoyIkcB4YFJmjgAmFe8BxgAjim0ccEqjX8lwkqSq687GtqXIzJmZeUfx+h/AA8AwYHfgzOKwM4E9ite7A2dlzWRgcESs28hX8p6TJFXdCrjnFBEbAVsBtwBDM3NmsespYGjxehjwRN3HphdtM1lGVk6SVHUNDutFxLiIuK1uG7e400fEQOBC4JDMfOG1l86kCc/itXKSpKprsHLKzAnAhJ6OiYiVqAXT2Zl5UdH8t4hYNzNnFsN2s4r2GcAGdR9fv2hbZlZOklR1TbrnFBEBnAY8kJnH1+2aCOxbvN4XuKSu/fPFrL1RwOy64b9lYuUkSVWXTfsV7vbA54B7IuKuou2bwNHABRExFngM+FSx73JgV2AqMAfYv9ELG06SVHVNmhCRmTcCsYTdOy/m+AQOXB7XNpwkqep6MURXNYaTJFVdGy5fZDhJUtUZTpKk0jGcJEml091+z8wwnCSp6qycJEmlYzhJkkrHqeSSpLLJ5q0Q0TKurSdJKh0rJ0mqOof1JEml44QISVLp+DsnSVLpWDlJksomrZwkSaVj5SRJKh1n60mSSqcNf4RrOElSxaWVkySpdKycJEllY+UkSSqfNqycIttwCmJfFhHjMnNCq/uhvsP/59QMrkrefsa1ugPqc/x/Tsud4SRJKh3DSZJUOoZT+3HsXyua/89puXNChCSpdKycJEmlYzi1kYgYHREPRsTUiBjf6v6ofUXE6RExKyLubXVf1J4MpzYREZ3AScAYYCSwd0SMbG2v1MbOAEa3uhNqX4ZT+9gWmJqZ0zJzHnAesHuL+6Q2lZnXA8+2uh9qX4ZT+xgGPFH3fnrRJkmVYzhJkkrHcGofM4AN6t6vX7RJUuUYTu3jVmBERAyPiP7AXsDEFvdJkhpiOLWJzFwAHARcCTwAXJCZ97W2V2pXEXEucDOwaURMj4ixre6T2osrREiSSsfKSZJUOoaTJKl0DCdJUukYTpKk0jGcJEmlYzipUiKiKyLuioh7I+LXEbHqGzjXGRGxZ/H61J4Wyo2IHSNiuwau8WhErNVoH6W+ynBS1byUme/KzHcA84AD6ndGRL9GTpqZX8zM+3s4ZEdgmcNJUmMMJ1XZDcDGRVVzQ0RMBO6PiM6IODYibo2IuyPi3wCi5v8Xz7z6A7D2whNFxLURsU3xenRE3BERf46ISRGxEbUQPLSo2t4XEW+OiAuLa9waEdsXnx0SEVdFxH0RcSoQK/Y/idQeGvpbptRqRYU0Bvh90bQ18I7MfCQixgGzM/M9ETEA+FNEXAVsBWxK7XlXQ4H7gdMXOe+bgZ8DOxTnWjMzn42InwIvZuZxxXHnACdk5o0RsSG1lTk2A44CbszM70TEboArJ0gNMJxUNatExF3F6xuA06gNt03JzEeK9l2ALRbeTwIGASOAHYBzM7MLeDIi/riY848Crl94rsxc0jOLPgiMjHilMHpTRAwsrvHx4rOXRcRzDX5PqU8znFQ1L2Xmu+obioD4Z30TcHBmXrnIcbsux350AKMyc+5i+iLpDfKek9rRlcCXI2IlgIjYJCJWA64HPl3ck1oX+MBiPjsZ2CEihhefXbNo/wewet1xVwEHL3wTEQsD83rgM0XbGGCN5fatpD7EcFI7OpXa/aQ7IuJe4GfURgkuBh4u9p1FbVXt18jMp4FxwEUR8Wfg/GLXpcD/WzghAvgKsE0x4eJ+Xp01+G1q4XYfteG9x5v0HaW25qrkkqTSsXKSJJWO4SRJKh3DSZJUOoaTJKl0DCdJUukYTpKk0jGcJEmlYzhJkkrn/wDMBpvhUbvRmQAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test2)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "GDFw1XJCFkjR"
   },
   "source": [
    "## Models 3 & 4: Additional layers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "E7oxJGWhFkjo"
   },
   "source": [
    "### Building"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "id": "cXIP6INbmr_A"
   },
   "outputs": [],
   "source": [
    "backend.clear_session()\n",
    "#Setting the seed for random number generators\n",
    "np.random.seed(42)\n",
    "import random\n",
    "random.seed(42)\n",
    "tf.random.set_seed(42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "id": "giM8fS5HFkjp"
   },
   "outputs": [],
   "source": [
    "#select layers within the neural network\n",
    "model_2 = Sequential()\n",
    "model_2.add(Dense(12, activation='relu',input_shape=(x_train1.shape[1],))) #input layer\n",
    "model_2.add(Dense(9, activation='relu')) #hidden layer\n",
    "model_2.add(Dense(6, activation='relu')) #hidden layer\n",
    "model_2.add(Dense(3, activation='relu')) #hidden layer\n",
    "model_2.add(Dense(1, activation='sigmoid')) #output layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "id": "iX2Ahq1oFkjp"
   },
   "outputs": [],
   "source": [
    "#select optimizer, loss function, and target metric\n",
    "adam = Adam(learning_rate=1e-3)\n",
    "model_2.compile(loss='binary_crossentropy', optimizer=adam, metrics=metric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "auf6JNl8Fkjq",
    "outputId": "00a262a1-3e79-403d-afc8-4a10ae59e09d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " dense (Dense)               (None, 12)                156       \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 9)                 117       \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 6)                 60        \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 3)                 21        \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 1)                 4         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 358\n",
      "Trainable params: 358\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#provide summary of model\n",
    "model_2.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8a-iLRGgFkjr"
   },
   "source": [
    "### Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "id": "AcPlJGg3Fkjr"
   },
   "outputs": [],
   "source": [
    "#train model across select number of epochs\n",
    "epochs=30\n",
    "history = model_2.fit(x_train1, y_train1, validation_split=0.2, epochs=epochs, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 350
    },
    "id": "J1YUDIE2Fkjs",
    "outputId": "8b8102b8-1b7e-4586-a9d6-3e3822712978"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot progression of losses across training epochs\n",
    "N = epochs\n",
    "plt.figure(figsize=(7,5))\n",
    "plt.plot(np.arange(0, N), history.history[\"loss\"], label=\"training\")\n",
    "plt.plot(np.arange(0, N), history.history[\"val_loss\"], label=\"validation\")\n",
    "\n",
    "plt.title(\"Training and validation losses across epochs\")\n",
    "plt.xlabel(\"Epoch\")\n",
    "plt.ylabel(\"Losses\")\n",
    "plt.legend(loc=\"best\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 294
    },
    "id": "0zVoR6DFFkjs",
    "outputId": "4ddcb8eb-decc-45f1-e2ba-20d8648b2b8d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot progression of accuracy across training epochs\n",
    "loss, accuracy  = model_2.evaluate(x_train1, y_train1, verbose=False)\n",
    "    \n",
    "plt.plot(history.history['auc'])\n",
    "plt.plot(history.history['val_auc'])\n",
    "plt.title('Model AUC across epochs')\n",
    "plt.ylabel('AUC')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['training', 'validation'], loc='best')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XNu_0dY-tGwi"
   },
   "source": [
    "- The ANN trained appears to minimally overfit the training data and to have reached a steady state of performance prior to the final epoch."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_n_tdkuwFkjt"
   },
   "source": [
    "### Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 350
    },
    "id": "4OXBcPsyFkjt",
    "outputId": "8be39568-ed57-4ee0-c023-39277c0f9f67"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot ROC curve on testing data\n",
    "roc_auc_test_2 = roc_auc_score(y_test1, model_2.predict(x_test1))\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_2.predict(x_test1))\n",
    "plt.figure(figsize=(7, 5))\n",
    "plt.plot(fpr, tpr, label=\"Model_1 (area = %0.2f)\" % roc_auc_test_2)\n",
    "plt.plot([0, 1], [0, 1], \"r--\")\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.01])\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate\")\n",
    "plt.title(\"Receiver operating characteristic\")\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "DV0utr5DFkju",
    "outputId": "29d391a5-f84b-4775-ae53-2fc025ffdc7c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.88      0.96      0.92      1585\n",
      "           1       0.76      0.48      0.59       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.82      0.72      0.75      2000\n",
      "weighted avg       0.85      0.86      0.85      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set baseline threshold\n",
    "threshold = 0.5\n",
    "#predict the labels of training data\n",
    "y_pred_test3=np.round(model_2.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(y_test1,y_pred_test3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "xERZlt6vFkju",
    "outputId": "ab25dfb0-f78d-4dad-f20b-a6ca5a298650"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test3)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xHXNO-UyFkjv"
   },
   "source": [
    "### Optimization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "tp7q8sMnFkjw",
    "outputId": "8265cb52-09b4-4864-80bc-64a78ca374f5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.2471463\n"
     ]
    }
   ],
   "source": [
    "#optimal threshold as per AUC-ROC curve\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_2.predict(x_test1))\n",
    "\n",
    "optimal_idx = np.argmax(tpr - fpr)\n",
    "optimal_threshold_auc_roc = thresholds[optimal_idx]\n",
    "print(optimal_threshold_auc_roc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "oUDeDJXdFkjw",
    "outputId": "950825a1-b360-45fd-ace1-1bc9c877668c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      0.85      0.89      1585\n",
      "           1       0.56      0.73      0.64       415\n",
      "\n",
      "    accuracy                           0.83      2000\n",
      "   macro avg       0.74      0.79      0.76      2000\n",
      "weighted avg       0.85      0.83      0.83      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set optimal threshold\n",
    "threshold = optimal_threshold_auc_roc\n",
    "#predict the labels of training data\n",
    "y_pred_test4=np.round(model_2.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "print(classification_report(y_test1,y_pred_test4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "akWmkHSHFkjw",
    "outputId": "0ad3157f-7ab7-40f4-924b-e5cbc8c3f24d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test4)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "wTgycpUXGsnF"
   },
   "source": [
    "## Models 5 & 6: Incorporate batch normalization and dropout layers "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "AGmfXiZcGsnR"
   },
   "source": [
    "### Building"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "id": "apKh73_5mu0b"
   },
   "outputs": [],
   "source": [
    "backend.clear_session()\n",
    "#Setting the seed for random number generators\n",
    "np.random.seed(42)\n",
    "import random\n",
    "random.seed(42)\n",
    "tf.random.set_seed(42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "id": "AAjTScipGsnR"
   },
   "outputs": [],
   "source": [
    "#select layers within the neural network\n",
    "model_3 = Sequential()\n",
    "model_3.add(Dense(12, activation='relu',input_shape=(x_train1.shape[1],))) #input layer\n",
    "model_3.add(BatchNormalization()) #batch normalization\n",
    "model_3.add(Dense(9, activation='relu')) #hidden layer\n",
    "model_3.add(Dropout(0.1)) #dropout 10% of training data\n",
    "model_3.add(BatchNormalization()) #batch normalization\n",
    "model_3.add(Dense(6, activation='relu')) #hidden layer\n",
    "model_3.add(BatchNormalization()) #batch normalization\n",
    "model_3.add(Dropout(0.1)) #dropout 10% of training data\n",
    "model_3.add(Dense(3, activation='relu')) #hidden layer\n",
    "model_3.add(Dense(1, activation='sigmoid')) #output layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "id": "U_-qCTt6GsnR"
   },
   "outputs": [],
   "source": [
    "#select optimizer, loss function, and target metric\n",
    "adam = Adam(learning_rate=1e-3)\n",
    "model_3.compile(loss='binary_crossentropy', optimizer=adam, metrics=metric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "t4O59iFrGsnR",
    "outputId": "a34d6e9a-4288-48f7-f7f7-c3f18c5b68e2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " dense (Dense)               (None, 12)                156       \n",
      "                                                                 \n",
      " batch_normalization (BatchN  (None, 12)               48        \n",
      " ormalization)                                                   \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 9)                 117       \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 9)                 0         \n",
      "                                                                 \n",
      " batch_normalization_1 (Batc  (None, 9)                36        \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 6)                 60        \n",
      "                                                                 \n",
      " batch_normalization_2 (Batc  (None, 6)                24        \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 6)                 0         \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 3)                 21        \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 1)                 4         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 466\n",
      "Trainable params: 412\n",
      "Non-trainable params: 54\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#provide summary of model\n",
    "model_3.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "T_4I1y2nGsnR"
   },
   "source": [
    "### Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "id": "wVZ3CI0nGsnR"
   },
   "outputs": [],
   "source": [
    "#train model across select number of epochs\n",
    "epochs=30\n",
    "history = model_3.fit(x_train1, y_train1, validation_split=0.2, epochs=epochs, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 350
    },
    "id": "m453LkWzGsnS",
    "outputId": "6534be93-c3ba-4fdd-e1d1-2ec3ae1010f9"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot progression of losses across training epochs\n",
    "N = epochs\n",
    "plt.figure(figsize=(7,5))\n",
    "plt.plot(np.arange(0, N), history.history[\"loss\"], label=\"training\")\n",
    "plt.plot(np.arange(0, N), history.history[\"val_loss\"], label=\"validation\")\n",
    "\n",
    "plt.title(\"Training and validation losses across epochs\")\n",
    "plt.xlabel(\"Epoch\")\n",
    "plt.ylabel(\"Losses\")\n",
    "plt.legend(loc=\"best\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 294
    },
    "id": "i_pFWXhEGsnS",
    "outputId": "8f1e9e0a-4204-46b0-ad96-61dfae01d82d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot progression of accuracy across training epochs\n",
    "loss, accuracy  = model_3.evaluate(x_train1, y_train1, verbose=False)\n",
    "    \n",
    "plt.plot(history.history['auc'])\n",
    "plt.plot(history.history['val_auc'])\n",
    "plt.title('Model AUC across epochs')\n",
    "plt.ylabel('AUC')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['training', 'validation'], loc='best')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "affP13Jvtgef"
   },
   "source": [
    "- The ANN trained appears to minimally overfit the training data and to have reached a steady state of performance prior to the final epoch.\n",
    "- Additionally, the ANN appears to perform consistently better on the validation data than the training data across the epochs."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8KDJmbfJGsnS"
   },
   "source": [
    "### Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 350
    },
    "id": "JaKwRuMKGsnS",
    "outputId": "c3a95762-c022-42ed-f3a4-73797a9b7a12"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot ROC curve on testing data\n",
    "roc_auc_test_3 = roc_auc_score(y_test1, model_3.predict(x_test1))\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_3.predict(x_test1))\n",
    "plt.figure(figsize=(7, 5))\n",
    "plt.plot(fpr, tpr, label=\"Model_1 (area = %0.2f)\" % roc_auc_test_3)\n",
    "plt.plot([0, 1], [0, 1], \"r--\")\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.01])\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate\")\n",
    "plt.title(\"Receiver operating characteristic\")\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "a74JSoS3GsnS",
    "outputId": "d222f6da-9cd7-4e54-8895-5e1022760110"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.97      0.92      1585\n",
      "           1       0.79      0.44      0.57       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.83      0.71      0.74      2000\n",
      "weighted avg       0.85      0.86      0.84      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set baseline threshold\n",
    "threshold = 0.5\n",
    "#predict the labels of training data\n",
    "y_pred_test5=np.round(model_3.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(y_test1,y_pred_test5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "JqXVuYzbGsnS",
    "outputId": "73d99f2d-2bbd-4b9d-ee64-58b6dc230402"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAacAAAE9CAYAAACvEq29AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAc3UlEQVR4nO3de7RVVf338fcXkZBEAS+IgD/MUFPTNFPMMu/XEu2npvYkGXWqYZZppVn9/HkrTcsnrZ9FQWKppWaPlJbyeEm7mLeUUEt4oARC0VA0r3HO9/ljL3CDXA7bs89Za5/3a4w13Huuufea28Hgw3fOudeOzESSpDLp09MDkCRpeYaTJKl0DCdJUukYTpKk0jGcJEmlYzhJkkqnb08PYGU6Ht/SPe7qNgdsukNPD0G9zNSOa6Kr3qvRvy/7bPJol42hq1k5SZJKp7SVkySpczroaOh1Za5ODCdJqrj2bCycyhwAZR6bJKkTOmi9JXrDSZIqrtFpvTIznCSp4tpb8AbehpMkVZzTepKk0mk3nCRJZWPlJEkqHdecJEml03p79QwnSao815wkSaXT3nrZZDhJUtU5rSdJKp12SvvLFw0znCSp4jqc1pMklU0rVk5l/jkPSVIvZeUkSRXXipWT4SRJFdeRrRdOTutJUsW1Ew0dqxMRkyJiQURMX8G5UyIiI2LD4nlExMURMTMipkXETnV9x0XEjOIY15nPZDhJUsW106ehoxMuAw5cvjEiRgL7A4/VNR8EjC6ONuDSou8Q4AxgV2AX4IyIGLy6CxtOklRxHRkNHauTmXcAC1dw6iLgC7DMfZPGApdnzV3AoIgYBhwATM3MhZn5NDCVFQTe8lxzkqSK684NERExFpiXmQ9GLHPd4cCcuudzi7aVta+S4SRJFdeejU2CRUQbtSm4JSZk5oRV9B8AnE5tSq+pDCdJqriOBldoiiBaaRitwBbA5sCSqmkEcH9E7ALMA0bW9R1RtM0D9lyu/fbVXcg1J0mquGbt1lteZv45MzfOzFGZOYraFN1Omfk4MAU4rti1NwZYlJnzgZuA/SNicLERYv+ibZWsnCSp4hqd1ludiLiKWtWzYUTMBc7IzIkr6X4jcDAwE3gBOB4gMxdGxNnAPUW/szJzRZsslmE4SVLFdTRpQ0RmHrOa86PqHidwwkr6TQImrcm1DSdJqrhOfmepUgwnSaq4Zk3r9STDSZIqrtHdemVmOElSxbW34I1fDSdJqrhWXHNqvU8kSao8KydJqrgON0RIksqmFaf1DCdJqjg3REiSSset5JKk0vFLuJKk0mnWvfV6kuEkSRVn5SRJKh1360mSSqfD3XqSpLKxcpIklY53iJAklU67u/UkSWVj5SRJKh0rJ0lS6Vg5SZJKpxW/hNt6n0iSVHlWTpJUcd5bT5JUOq04rWc4SVLFefsiSVLpePsiSVLptGLl1HpxK0m9TAd9GjpWJyImRcSCiJhe13ZBRPwlIqZFxM8jYlDduS9GxMyI+GtEHFDXfmDRNjMiTuvMZzKcJKni2jMaOjrhMuDA5dqmAttl5vbAo8AXASJiG+BoYNviNf8TEWtFxFrAd4CDgG2AY4q+q+S0niRVXLOm9TLzjogYtVzbzXVP7wKOKB6PBX6SmS8DsyNiJrBLcW5mZs4CiIifFH0fXtW1rZwkqeI6sk9DR0S0RcS9dUfbGl76I8CvisfDgTl15+YWbStrXyUrJ0mquEZv/JqZE4AJjbw2Ir4ELAauaOjiq2E4VcCXzoPb/wBDBsMvLqu1ffuHcM0vYUixFHnSx+A9Y2DaI3DGhbW2TDjhw7DfHrXnzz4HX7kAZsyGAM45FXbcrps/jCqvT58+fOee83hq3kK+cuh5vG2v7Wi74EP07deXGffN4hsfvZSO9o6eHmav0t279SLiw8B7gX0yM4vmecDIum4jijZW0b5ShlMFHHYQHPt+OO2ry7aPOxI+cvSybaM3h2u+B337woJ/wuEfgb3eWXv+1UvgXbvAt86CV/4NL73UfZ9BrePwzxzMY4/MY8B66xARfP6yE/jCvmcxb8Z8xp35AfYftye/nnRrTw+zV+nOu5JHxIHAF4D3ZOYLdaemAFdGxDeBTYHRwN3U/i08OiI2pxZKRwPHru46rjlVwDt2gEEDO9d3nf61IAJ45RWI4h9Uz/0L7n0Qjjik9rzf2rBeJ99TWmLD4UPY9eCd+NXEWwBYb4OBLH5lMfNmzAfgvqkP8u7379qTQ+yVOoiGjtWJiKuAPwBbRcTciBgPfBsYCEyNiAci4rsAmfkQcDW1jQ6/Bk7IzPbMXAx8CrgJeAS4uui7Sk2rnCJia2o7MpYsfM0DpmTmI826Zm9zxc/h+ptgu63gCyfA+kXYPPgwfOl8mP8EnHd6Lazmzq9NAZ5+Hvx1JmyzFZx+IgxYp2c/g6rlkxcdz/dP/THrDOwPwKKnnmWtvmux5dvfxKP3zWKPI3Zjo5Eb9vAoe59ObgtfY5l5zAqaJ66i/7nAuStovxG4cU2u3ZTKKSJOBX5CrZy7m1dLu6s6+wUsrdrRY+HmK+HnE2GjDeDr33n13A7bwC8nw9Xfhe9fAS+/DO3t8PCM2uuumwgD+sP3r+y58at6dj1kJ555chEz7p+1TPu5x/xvPvHND3PJXV/jhededL2pBzS6W6/MmlU5jQe2zcx/1zcWc5EPAeet6EXFNsY2gEu/vjFtH1q/ScOrvg2HvPr4yPfCJ7742j5bjKpVRjNmw9CNascOxVff9n+P4aQ1s+3uW7Pb+3Zml4N2pF//fgxYbx1OvfxEzj/uEk5+z38B8Pb9tmfE6GE9PNLepxVvX9SscOqgtiD29+XahxXnVqh+W2PH41vmyvqpttlh4w1qj6feWdsIAbXpu002qk3lzXscZj0GwzeBwYNg2EYw+zHYfDO4635486geG74qaNLpVzLp9Nq/aLZ/zzYcecqhnH/cJQzaaD2eefJZ1u7Xlw984TCu/Op1PTxStYJmhdNJwC0RMYNXv3y1GfBmagtjWgOnnAl3PwDPLII9j4BPHQ93/wn+MrO24WH4JvDfn6v1vW9arSJau2/t3H99thZMAF/6DHz+HPj3v2HkpnCuE6zqAkd+fixjDtmJ6NOHX3z3Jh64bfrqX6Qu1Yo/NhivblHv4jeO6EPt1hX1GyLuycz2zrzeyknd6YBNd+jpIaiXmdpxTZclyjF3tTX09+VVYyaUNtWatlsvMzuo3XdJktREZd/c0Ai/hCtJFeeGCElS6bTimpPhJEkVZ+UkSSodw0mSVDqGkySpdAwnSVLpuCFCklQ6Vk6SpNIxnCRJpWM4SZJKx3CSJJVOGk6SpLJxt54kqXRacVqv9e6zLkmqPCsnSao415wkSaXTitN6hpMkVZyVkySpdKycJEmlk9nTI+h6hpMkVVwrfs/JreSSVHGZ0dCxOhExKSIWRMT0urYhETE1ImYU/x1ctEdEXBwRMyNiWkTsVPeacUX/GRExrjOfyXCSpIrryGjo6ITLgAOXazsNuCUzRwO3FM8BDgJGF0cbcCnUwgw4A9gV2AU4Y0mgrYrhJEkVl9nYsfr3zTuAhcs1jwUmF48nA4fVtV+eNXcBgyJiGHAAMDUzF2bm08BUXht4r+GakyRVXDdvJR+amfOLx48DQ4vHw4E5df3mFm0ra18lKydJqrhG15wioi0i7q072tbsuplAU/YKWjlJUsU1+j2nzJwATFjDlz0REcMyc34xbbegaJ8HjKzrN6JomwfsuVz77au7iJWTJFVcs9acVmIKsGTH3Tjg+rr244pde2OARcX0303A/hExuNgIsX/RtkpWTpJUcc1ac4qIq6hVPRtGxFxqu+7OA66OiPHA34Gjiu43AgcDM4EXgONrY8uFEXE2cE/R76zMXH6TxWsYTpJUcc0Kp8w8ZiWn9llB3wROWMn7TAImrcm1DSdJqrgWvHuRa06SpPKxcpKkivMnMyRJ5dOC83qGkyRVnJWTJKl0/D0nSVLpWDlJksrHcJIklY3TepKk8jGcJEll45qTJKl8rJwkSWVj5SRJKh8rJ0lS+Vg5SZLKxspJklQ6hpMkqXRacEOEPzYoSSodKydJqjhvXyRJKh/DSZJUOi245tSpcIqIdwKj6vtn5uVNGpMkaQ1Eb6ycIuJHwBbAA0B70ZyA4SRJZdAbwwnYGdgmsxWX3CSpBbTgtF5ntpJPBzZp9kAkSQ3KBo8SW2nlFBG/oDb8gcDDEXE38PKS85l5aPOHJ0larZIHTSNWNa13YbeNQpLUuBYMp5VO62XmbzLzN8DBSx7Xt3XfECVJq5TR2LEaEfHZiHgoIqZHxFUR0T8iNo+IP0bEzIj4aUT0K/q+oXg+szg/6vV8pM6sOe23graDXs9FJUldJ7KxY5XvGTEc+DSwc2ZuB6wFHA2cD1yUmW8GngbGFy8ZDzxdtF9U9GvYSsMpIj4ZEX8Gto6IaXXHbODPr+eikqQu1LwNEX2BdSKiLzAAmA/sDVxbnJ8MHFY8Hls8pzi/T0Q0vI1wVWtOVwK/Ar4GnFbX/lxmLmz0gpKk8svMeRFxIfAY8CJwM3Af8ExmLi66zQWGF4+HA3OK1y6OiEXABsBTjVx/peGUmYuARRFx6nKn1o2IdTPzsUYu2FkH73dkM99eWkbfUS+vvpNUUo3eISIi2oC2uqYJmTmhODeYWjW0OfAMcA1w4Osa6BrozJdwb6BWAAbQn9pA/wps28RxSZI6q8Ev4RZBNGElp/cFZmfmkwARcR2wOzAoIvoW1dMIYF7Rfx4wEphbTAOuD/yzoYHRiQ0RmfnWzNy++O9oYBfgD41eUJLUxZqz5vQYMCYiBhRrR/sADwO3AUcUfcYB1xePpxTPKc7f+nruLLTGPzaYmfcDuzZ6QUlS+WXmH6ltbLif2ia4PtSqrFOBkyNiJrU1pYnFSyYCGxTtJ7PsXoU11pkbv55c97QPsBPwj9dzUUlSF2rSl3Az8wzgjOWaZ1GbQVu+70tAl20W6Mya08C6x4uprUH9rKsGIEl6fXrdT2ZExFrAwMz8XDeNR5K0pnpTOC3ZjRERu3fngCRJa6g3hRNwN7X1pQciYgq1Pe7PLzmZmdc1eWySpE7oddN6hf7U9qrvzavfd0rAcJKkMmjBHxtcVThtXOzUm86robREC+a0JFVUC/6NvKpwWgtYl2VDaYkW/F8hSdXU26b15mfmWd02EklSY3pZOLXeJKYktaDeVjnt022jkCQ1rjeFk7/ZJEkV0ZvCSZJUDa04rbfGdyWXJKnZrJwkqepasHIynCSp4pzWkySpG1g5SVLVtWDlZDhJUtUZTpKksmnFNSfDSZKqznCSJJWNlZMkqXwMJ0lS6RhOkqSycVpPklQ+hpMkqXQMJ0lS2bTitJ731pOkqssGj06IiEERcW1E/CUiHomI3SJiSERMjYgZxX8HF30jIi6OiJkRMS0idmr0IxlOklRxkY0dnfQt4NeZuTWwA/AIcBpwS2aOBm4pngMcBIwujjbg0kY/k+EkSVXXpMopItYH9gAmAmTmK5n5DDAWmFx0mwwcVjweC1yeNXcBgyJiWCMfyXCSpKpr3rTe5sCTwA8j4k8R8YOIeCMwNDPnF30eB4YWj4cDc+peP7doW2OGkyRVXDR6RLRFxL11R9tyb90X2Am4NDN3BJ7n1Sk8ADJzDVawOs/depJUdQ1GQ2ZOACasostcYG5m/rF4fi21cHoiIoZl5vxi2m5BcX4eMLLu9SOKtjVm5SRJWqHMfByYExFbFU37AA8DU4BxRds44Pri8RTguGLX3hhgUd303xqxcpKkimvy95xOBK6IiH7ALOB4aoXN1RExHvg7cFTR90bgYGAm8ELRtyGGkyRVXRPDKTMfAHZewal9VtA3gRO64rqGkyRVXQveIcJwkqSKa8XbFxlOklR1hpMkqWysnCRJ5WM4SZLKxspJklQ+hpMkqXQMJ0lS2TitJ0kqH8NJklQ2ka2XToaTJFVd62WT4SRJVeeakySpfFownPyxQUlS6Vg5SVLFOa0nSSofw0mSVDZWTpKk8jGcJEllY+UkSSof7xAhSSobKydJUvkYTuppGw5dn89/9QgGbbAuZHLjtfdw/RV/4LhP7ctue72Fjo7kmYX/4htf/hkLn3yOvQ7ZgaM+sgcEvPj8y1xy9hRmP/p4T38MVcRnzz+KXfbahmf++S8+edCFALzpLZty4jn/ydpv6Et7ewff+cp1PDptDmP23ZbjTj6Ajo6kvb2DCWdfz0P3/q1nP0AvER09PYKuF1nSucoD3/qlcg6shw3ZcCBDNhrIzEf+wToD+nHJT0/grM/8mKeeeJYXnn8ZgLHH7sZmW2zMJWdfz1t22Iw5sxfwr2dfYud3bcn/+uTenPTB7/bwpyifKP7faVnbveNNvPjCy3zuwmOWhtO5kz/Gzyfdyb2/+Qvv2HNrjmjbi1OPvZT+A/rx0guvADBq62GcfsmHaNvv6z05/FL71awLo6ve651HfaOhvy9/f/UpXTaGrmblVDELn3qOhU89B8CLL7zCnNlPssHQ9Xhs1pNL+/RfZ22W/KPjkQcfW9r+l2mPseHQ9bt3wKq06ffMYuPhg5dpy4QB674BgAED+/PPBYsAlgYTQP91+lHWf/i2ItecVCpDNx3EFlsP46/T5gIw7sT92PfQt/H8cy9z6vgfvKb/AYfvzL2/fbS7h6kW872zr+ecyR/jo198H9EnOOWIby899879t+PDnz+YQRusy3+Nn9iDo+xlWvAfAt1+49eIOL67r9mK+q/Tjy9fdCzfO/+GpdN5ky+Zyof2u4DbbniA9x2z2zL9t3/H5hzw/rcz8aJf98Rw1UIO+eBuTDhnCse96xwmnDOFk84/cum53988nbb9vs5ZH7+M404+oAdH2btENnaUWU/clfzMlZ2IiLaIuDci7p2z8E/dOaZKWatvH75y0bHcdsOD/O6Wh19z/tYbHuRd+2679PnmWw7lpDMP58xP/5jnFr3YnUNVC9r3P3fmd7/+MwB33vggW22/2Wv6TL9nFpuM3ID1Bg/o7uH1Ttng0QkRsVZE/Ckiflk83zwi/hgRMyPipxHRr2h/Q/F8ZnF+1Ov5SE0Jp4iYtpLjz8DQlb0uMydk5s6ZufPIITs2Y2gt4bNnvp/HZi3gust/t7Rt0802WPp4t73fwpzZtTWojTZZn69c9EEu+OK1zPv7P7t9rGo9/3ziWd666xYAvO2db2be354CYNh/vPpncItth7N2v748+/QLPTLG3qbJldNngEfqnp8PXJSZbwaeBsYX7eOBp4v2i4p+DWvWmtNQ4ABqA68XwO+bdM1eYdsd/4N9D92R2Y8+zneu+RQAl118Mwcc/nZGjNqIzOSJfzzDJWdfD8AHP7E3AwcN4FNfPhSA9vYOPn30//TY+FUtp37rg2y/6xasN/iN/Oh3X+ZH37qZi0+/ho9/5TDW6tuHV15ezMVfugaAdx24Pfsc/nYWL27nlZf+zXmf/lEPj74XadKaU0SMAA4BzgVOjogA9gaOLbpMBv4buBQYWzwGuBb4dkRENrgzpilbySNiIvDDzPztCs5dmZnHruBly3ArubqTW8nV3bpyK/m7D7ugob8v7/w/n1/lGCLiWuBrwEDgc8CHgbuK6oiIGAn8KjO3i4jpwIGZObc49/+AXTPzqUbG1pRpvcwcv6JgKs6tNpgkSZ3X6LRe/Tp/cbQtfc+I9wILMvO+nvhMbiWXpKprcJ4pMycAE1Zyenfg0Ig4GOgPrAd8CxgUEX0zczEwAphX9J8HjATmRkRfYH2g4YXuntitJ0nqQs3YEJGZX8zMEZk5CjgauDUzPwjcBhxRdBsHXF88nlI8pzh/a6PrTWDlJEnV19GtS/SnAj+JiHOAPwFLvm09EfhRRMwEFlILtIYZTpJUdU3Opsy8Hbi9eDwL2GUFfV4Cjly+vVGGkyRVXNnv9tAIw0mSqq4F761nOElSxVk5SZLKx3CSJJVNOK0nSSqdFvyZdsNJkirOykmSVD6tl02GkyRVnpWTJKls3EouSSqfFqycvCu5JKl0rJwkqeLCreSSpNJpwWk9w0mSqq71sslwkqSq80u4kqTyMZwkSaXjhghJUtk4rSdJKh/DSZJUOoaTJKl0XHOSJJWNa06SpPIxnCRJpWM4SZJKx3CSJJVOC26I8PecJKniIrOhY7XvGzEyIm6LiIcj4qGI+EzRPiQipkbEjOK/g4v2iIiLI2JmREyLiJ0a/UyGkyRpZRYDp2TmNsAY4ISI2AY4DbglM0cDtxTPAQ4CRhdHG3Bpoxc2nCSp6jIbO1b7tjk/M+8vHj8HPAIMB8YCk4tuk4HDisdjgcuz5i5gUEQMa+QjueYkSVXX0fwNERExCtgR+CMwNDPnF6ceB4YWj4cDc+peNrdom88asnKSpKprsHKKiLaIuLfuaFvR20fEusDPgJMy89llL51JE37u0MpJkqquwa3kmTkBmLCqPhGxNrVguiIzryuan4iIYZk5v5i2W1C0zwNG1r18RNG2xqycJKnqmrTmFBEBTAQeycxv1p2aAowrHo8Drq9rP67YtTcGWFQ3/bdGrJwkqeqat+a0O/Ah4M8R8UDRdjpwHnB1RIwH/g4cVZy7ETgYmAm8ABzf6IUNJ0mqumzOt3Az87dArOT0Pivon8AJXXFtw0mSqs7bF0mSSqcbtpJ3N8NJkqrOykmSVDqGkySpdAwnSVLpdLTeb2YYTpJUdVZOkqTSMZwkSaXjVnJJUtlkk+4Q0ZO88askqXSsnCSp6pzWkySVjhsiJEml4/ecJEmlY+UkSSqbtHKSJJWOlZMkqXTcrSdJKp0W/BKu4SRJFZdWTpKk0rFykiSVjZWTJKl8WrByimzBLYi9WUS0ZeaEnh6Heg//zKkZvCt562nr6QGo1/HPnLqc4SRJKh3DSZJUOoZT63HuX93NP3Pqcm6IkCSVjpWTJKl0DKcWEhEHRsRfI2JmRJzW0+NR64qISRGxICKm9/RY1JoMpxYREWsB3wEOArYBjomIbXp2VGphlwEH9vQg1LoMp9axCzAzM2dl5ivAT4CxPTwmtajMvANY2NPjUOsynFrHcGBO3fO5RZskVY7hJEkqHcOpdcwDRtY9H1G0SVLlGE6t4x5gdERsHhH9gKOBKT08JklqiOHUIjJzMfAp4CbgEeDqzHyoZ0elVhURVwF/ALaKiLkRMb6nx6TW4h0iJEmlY+UkSSodw0mSVDqGkySpdAwnSVLpGE6SpNIxnFQpEdEeEQ9ExPSIuCYiBryO97osIo4oHv9gVTfKjYg9I+KdDVzjbxGxYaNjlHorw0lV82Jmvi0ztwNeAT5RfzIi+jbyppn50cx8eBVd9gTWOJwkNcZwUpXdCby5qGrujIgpwMMRsVZEXBAR90TEtIj4OEDUfLv4zav/C2y85I0i4vaI2Ll4fGBE3B8RD0bELRExiloIfrao2t4dERtFxM+Ka9wTEbsXr90gIm6OiIci4gdAdO//Eqk1NPSvTKmnFRXSQcCvi6adgO0yc3ZEtAGLMvMdEfEG4HcRcTOwI7AVtd+7Ggo8DExa7n03Ar4P7FG815DMXBgR3wX+lZkXFv2uBC7KzN9GxGbU7szxFuAM4LeZeVZEHAJ45wSpAYaTqmadiHigeHwnMJHadNvdmTm7aN8f2H7JehKwPjAa2AO4KjPbgX9ExK0reP8xwB1L3iszV/abRfsC20QsLYzWi4h1i2u8v3jtDRHxdIOfU+rVDCdVzYuZ+bb6hiIgnq9vAk7MzJuW63dwF46jDzAmM19awVgkvU6uOakV3QR8MiLWBoiILSPijcAdwAeKNalhwF4reO1dwB4RsXnx2iFF+3PAwLp+NwMnLnkSEUsC8w7g2KLtIGBwl30qqRcxnNSKfkBtPen+iJgOfI/aLMHPgRnFucup3VV7GZn5JNAGXBcRDwI/LU79Ajh8yYYI4NPAzsWGi4d5ddfgmdTC7SFq03uPNekzSi3Nu5JLkkrHykmSVDqGkySpdAwnSVLpGE6SpNIxnCRJpWM4SZJKx3CSJJWO4SRJKp3/D27psRVRRC0HAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test5)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LFBaqW-5GsnS"
   },
   "source": [
    "### Optimization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "v8DPwfr8GsnT",
    "outputId": "f9e78646-611f-4c59-884c-c9ec60146abd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.21930292\n"
     ]
    }
   ],
   "source": [
    "#optimal threshold as per AUC-ROC curve\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_3.predict(x_test1))\n",
    "\n",
    "optimal_idx = np.argmax(tpr - fpr)\n",
    "optimal_threshold_auc_roc = thresholds[optimal_idx]\n",
    "print(optimal_threshold_auc_roc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "osW30w0uGsnT",
    "outputId": "6d863fa8-6a4d-4fd3-e784-15fe14635b5d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.83      0.88      1585\n",
      "           1       0.53      0.74      0.62       415\n",
      "\n",
      "    accuracy                           0.81      2000\n",
      "   macro avg       0.73      0.79      0.75      2000\n",
      "weighted avg       0.84      0.81      0.82      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set optimal threshold\n",
    "threshold = optimal_threshold_auc_roc\n",
    "#predict the labels of training data\n",
    "y_pred_test6=np.round(model_3.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "print(classification_report(y_test1,y_pred_test6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "q9xRzB_FGsnT",
    "outputId": "6c40486f-e5dd-42a2-fdb0-58f6b5b18ca5"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test6)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HzA22Pc3E2Kn"
   },
   "source": [
    "## Model 7 & 8: Randomized Search CV\n",
    "\n",
    "- To implement a Randomized Search, we need to define a model function that can take in the hyper parameters. Note, we will use a Wrapper Class to integrate a NN model with Randomized Search."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "id": "Xk6nUvgPS2s5"
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras.optimizers import RMSprop\n",
    "from tensorflow.keras.optimizers import SGD\n",
    "\n",
    "rms = RMSprop()\n",
    "sgd = SGD()\n",
    "adam1 = Adam()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "id": "eMfYQIouE2Kn"
   },
   "outputs": [],
   "source": [
    "# ------------------- L1 Regularizer-----------------------\n",
    "\n",
    "def build_model(n_hidden_layer=1, n_neurons=12, input_shape=X_train.shape[1], learning_rate=0.01, active=\"relu\", drop=0.1, optimizers=adam1):\n",
    "\n",
    "    # 1. Sequential model: create model; Initializes the model\n",
    "    model = Sequential()\n",
    "    \n",
    "    model.add(keras.Input(shape=(input_shape,))) #specify the shape of input Layer\n",
    "    \n",
    "    #---------------------The number of neurons in layer n+1 is 2/3 * Number of neurons in layer n---------------\n",
    "    \n",
    "    for layer in range(n_hidden_layer):\n",
    "        \n",
    "        model.add(Dense(n_neurons, activation=active, kernel_regularizer=tf.keras.regularizers.L1(0.01)))\n",
    "\n",
    "        if n_hidden_layer%2==0:\n",
    "          \n",
    "          model.add(Dropout(drop)) #Adding\n",
    "        \n",
    "        n_neurons=int(round(n_neurons*2/3))\n",
    "    \n",
    "    model.add(Dense(1, activation = 'sigmoid'))\n",
    "    \n",
    "    # Compile model\n",
    "    \n",
    "    optimizer=optimizers\n",
    "\n",
    "    model.compile(optimizer = optimizer, loss = 'binary_crossentropy', metrics=metric) \n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "id": "_f64-CkFE2Ko"
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras.wrappers.scikit_learn import KerasClassifier\n",
    " \n",
    "model_opt = KerasClassifier(build_fn=build_model, epochs=30, batch_size=32, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "N8iRMIV-E2Ko",
    "outputId": "2ea6a339-14df-4450-f6b6-5bc57746eb97"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3,\n",
       "                   estimator=<keras.wrappers.scikit_learn.KerasClassifier object at 0x7f41e80fded0>,\n",
       "                   n_iter=5,\n",
       "                   param_distributions={'active': ['relu', 'selu'],\n",
       "                                        'learning_rate': [0.001, 0.01],\n",
       "                                        'n_hidden_layer': [2, 3, 4, 5],\n",
       "                                        'n_neurons': [9, 12, 15, 18],\n",
       "                                        'optimizers': [<keras.optimizer_v2.adam.Adam object at 0x7f41562f5f50>,\n",
       "                                                       <keras.optimizer_v2.gradient_descent.SGD object at 0x7f41562f77d0>,\n",
       "                                                       <keras.optimizer_v2.rmsprop.RMSprop object at 0x7f41562f5210>]},\n",
       "                   random_state=42, scoring='f1')"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "\n",
    "param_distribution = {\n",
    "    \"n_hidden_layer\": [2,3,4,5],\n",
    "    \"n_neurons\": [9,12,15,18],\n",
    "    \"active\": ['relu', 'selu'],\n",
    "    \"learning_rate\": [0.001,0.01],\n",
    "    \"optimizers\":[adam1,sgd,rms]\n",
    "}\n",
    "\n",
    "#Randomized Search CV: K-fold cross-validation to determine the optimal hyperparameters\n",
    "\n",
    "random_search_cv = RandomizedSearchCV(model_opt, param_distribution, n_iter=5, cv=3, scoring='f1', random_state=42)\n",
    "\n",
    "random_search_cv.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "r5OSKIizvJE3",
    "outputId": "62bac494-a119-4bf0-f0f2-67619a690f72"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best hyperparameters: {'optimizers': <keras.optimizer_v2.rmsprop.RMSprop object at 0x7f41562f5210>, 'n_neurons': 18, 'n_hidden_layer': 5, 'learning_rate': 0.001, 'active': 'selu'}\n"
     ]
    }
   ],
   "source": [
    "print(\"Best hyperparameters: %s\" % (random_search_cv.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "id": "VK2eAAqIJ_LU"
   },
   "outputs": [],
   "source": [
    "model_tuned = random_search_cv.best_estimator_.model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OqJjnDmHcG-3"
   },
   "source": [
    "### Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 350
    },
    "id": "IX55NNVncG-4",
    "outputId": "54636d7f-d277-4f7c-ca32-e608ed16f05b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot ROC curve on testing data\n",
    "roc_auc_test_4 = roc_auc_score(y_test1, model_tuned.predict(x_test1))\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_tuned.predict(x_test1))\n",
    "plt.figure(figsize=(7, 5))\n",
    "plt.plot(fpr, tpr, label=\"Model_1 (area = %0.2f)\" % roc_auc_test_4)\n",
    "plt.plot([0, 1], [0, 1], \"r--\")\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.01])\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate\")\n",
    "plt.title(\"Receiver operating characteristic\")\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Ixm0gAyMaOyH",
    "outputId": "31250cab-3858-4c16-df30-d4d878fd3b30"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.97      0.92      1585\n",
      "           1       0.79      0.44      0.56       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.83      0.70      0.74      2000\n",
      "weighted avg       0.85      0.86      0.84      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set baseline threshold\n",
    "threshold = 0.5\n",
    "#predict the labels of training data\n",
    "y_pred_test7=np.round(model_tuned.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(y_test1,y_pred_test7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "87sYhIvRbdeF",
    "outputId": "af79d4cb-6307-4e6a-bb6a-48239085a1c4"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test7)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8SgPNhVybdeG"
   },
   "source": [
    "### Optimization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "pwrgTGazbdeG",
    "outputId": "70f1f971-2dcf-4e05-fd71-7f0f367b4ef4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.17216419\n"
     ]
    }
   ],
   "source": [
    "#optimal threshold as per AUC-ROC curve\n",
    "fpr, tpr, thresholds = roc_curve(y_test1, model_tuned.predict(x_test1))\n",
    "\n",
    "optimal_idx = np.argmax(tpr - fpr)\n",
    "optimal_threshold_auc_roc = thresholds[optimal_idx]\n",
    "print(optimal_threshold_auc_roc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "DMOkfOOrbdeH",
    "outputId": "6f98e450-3c1a-44c4-9775-cbfa5c1f7441"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.82      0.87      1585\n",
      "           1       0.52      0.75      0.61       415\n",
      "\n",
      "    accuracy                           0.80      2000\n",
      "   macro avg       0.72      0.78      0.74      2000\n",
      "weighted avg       0.84      0.80      0.81      2000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#set optimal threshold\n",
    "threshold = optimal_threshold_auc_roc\n",
    "#predict the labels of training data\n",
    "y_pred_test8=np.round(model_tuned.predict(x_test1)>threshold,0).astype(int)\n",
    "\n",
    "#generate classification report\n",
    "print(classification_report(y_test1,y_pred_test8))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "id": "zeJ1G66kbdeH",
    "outputId": "16f651bd-582c-4873-c4be-64230e31d494"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create confusion matrix\n",
    "cm=confusion_matrix(y_test1,y_pred_test8)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "sns.heatmap(cm,annot=True,fmt='d',cmap='viridis')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Truth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QWDu8jxZH4Xt"
   },
   "source": [
    "## Models summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "aBxlJ6b1LnWf",
    "outputId": "cc400d6f-a403-4373-ee25-a69208f8c6bf"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model 1\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.88      0.96      0.92      1585\n",
      "           1       0.75      0.50      0.60       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.81      0.73      0.76      2000\n",
      "weighted avg       0.85      0.86      0.85      2000\n",
      "\n",
      " \n",
      "Model 2\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.77      0.84      1585\n",
      "           1       0.47      0.79      0.59       415\n",
      "\n",
      "    accuracy                           0.78      2000\n",
      "   macro avg       0.70      0.78      0.72      2000\n",
      "weighted avg       0.84      0.78      0.79      2000\n",
      "\n",
      " \n",
      "Model 3\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.88      0.96      0.92      1585\n",
      "           1       0.76      0.48      0.59       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.82      0.72      0.75      2000\n",
      "weighted avg       0.85      0.86      0.85      2000\n",
      "\n",
      " \n",
      "Model 4\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      0.85      0.89      1585\n",
      "           1       0.56      0.73      0.64       415\n",
      "\n",
      "    accuracy                           0.83      2000\n",
      "   macro avg       0.74      0.79      0.76      2000\n",
      "weighted avg       0.85      0.83      0.83      2000\n",
      "\n",
      " \n",
      "Model 5\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.97      0.92      1585\n",
      "           1       0.79      0.44      0.57       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.83      0.71      0.74      2000\n",
      "weighted avg       0.85      0.86      0.84      2000\n",
      "\n",
      " \n",
      "Model 6\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.83      0.88      1585\n",
      "           1       0.53      0.74      0.62       415\n",
      "\n",
      "    accuracy                           0.81      2000\n",
      "   macro avg       0.73      0.79      0.75      2000\n",
      "weighted avg       0.84      0.81      0.82      2000\n",
      "\n",
      " \n",
      "Model 7\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.97      0.92      1585\n",
      "           1       0.79      0.44      0.56       415\n",
      "\n",
      "    accuracy                           0.86      2000\n",
      "   macro avg       0.83      0.70      0.74      2000\n",
      "weighted avg       0.85      0.86      0.84      2000\n",
      "\n",
      " \n",
      "Model 8\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.82      0.87      1585\n",
      "           1       0.52      0.75      0.61       415\n",
      "\n",
      "    accuracy                           0.80      2000\n",
      "   macro avg       0.72      0.78      0.74      2000\n",
      "weighted avg       0.84      0.80      0.81      2000\n",
      "\n",
      " \n"
     ]
    }
   ],
   "source": [
    "#list of predictions for all models generated\n",
    "y_pred_tests = [y_pred_test1,y_pred_test2,y_pred_test3,y_pred_test4,y_pred_test5,y_pred_test6,y_pred_test7,y_pred_test8]\n",
    "\n",
    "modelx = 1\n",
    "\n",
    "#generate classification reports\n",
    "for y_pred in y_pred_tests:\n",
    "  print(\"Model \"+str(modelx))\n",
    "  print(classification_report(y_test1,y_pred))\n",
    "  print(\" \")\n",
    "  modelx+=1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "tuLLl_WLH7vP"
   },
   "source": [
    "## Conclusions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aGlM3chRm6r1"
   },
   "source": [
    "**Criteria**\n",
    "- My opinion of the best overall model is determined by the highest weighted average F1 score. This model provides the best combination of precision and recall across both possible outcomes.\n",
    "\n",
    "- However, if the costs of attracting new customers outweights the costs of retaining an existing customer, then the best model would be the one which provides the highest F1 score for the prediciting an exiting customer (i.e., 1).\n",
    "\n",
    "**Recommendation**\n",
    "- Across all models with either a baseline (50%) threshold or the optimal threshold, I'd recommend the client select one of two models based on the preceding two rationales:\n",
    "  - Model 1: KISS principle\n",
    "    - The baseline model with a baseline threshold provides an equivalent F1 score, but, as a tie-breaker, also provides a minimally better score for exiting customers than Model 3.\n",
    "  - Model 4: Find me the needles\n",
    "    - With additional layers and an optimal threshold, this model provides the best capability of identifying an exiting customer amongst a population of all customers."
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Bank_Churn_Prediction.ipynb",
   "provenance": []
  },
  "interpreter": {
   "hash": "d305d341dcf765741621c3ba47f91396df1aaaf8d2e5682afd6c7693bb4c5810"
  },
  "kernelspec": {
   "display_name": "Python 3.10.2 64-bit",
   "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": "3.10.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}