credit-card-fraud-detection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# importing the packages\n",
    "import datetime\n",
    "import warnings\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from datetime import date\n",
    "import sklearn.exceptions\n",
    "from pylab import rcParams\n",
    "import plot_functions as pf\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import linear_model\n",
    "from sklearn.svm import OneClassSVM\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from sklearn.ensemble import IsolationForest\n",
    "from imblearn.under_sampling import NearMiss \n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.neighbors import LocalOutlierFactor\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import datasets, metrics, preprocessing, svm\n",
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "from sklearn.metrics import classification_report,accuracy_score\n",
    "warnings.filterwarnings(\"ignore\", category=sklearn.exceptions.UndefinedMetricWarning)\n",
    "\n",
    "\n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Loading the data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Acknowledgements\n",
    "Link to the data set is the kaggle link as follows:\n",
    "\n",
    "https://www.kaggle.com/kartik2112/fraud-detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# loading data\n",
    "data_train = pd.read_csv(\"fraudTrain.csv\")\n",
    "data_test = pd.read_csv(\"fraudTest.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Exploratory Data Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.1 Data information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1296675 entries, 0 to 1296674\n",
      "Data columns (total 23 columns):\n",
      " #   Column                 Non-Null Count    Dtype  \n",
      "---  ------                 --------------    -----  \n",
      " 0   Unnamed: 0             1296675 non-null  int64  \n",
      " 1   trans_date_trans_time  1296675 non-null  object \n",
      " 2   cc_num                 1296675 non-null  int64  \n",
      " 3   merchant               1296675 non-null  object \n",
      " 4   category               1296675 non-null  object \n",
      " 5   amt                    1296675 non-null  float64\n",
      " 6   first                  1296675 non-null  object \n",
      " 7   last                   1296675 non-null  object \n",
      " 8   gender                 1296675 non-null  object \n",
      " 9   street                 1296675 non-null  object \n",
      " 10  city                   1296675 non-null  object \n",
      " 11  state                  1296675 non-null  object \n",
      " 12  zip                    1296675 non-null  int64  \n",
      " 13  lat                    1296675 non-null  float64\n",
      " 14  long                   1296675 non-null  float64\n",
      " 15  city_pop               1296675 non-null  int64  \n",
      " 16  job                    1296675 non-null  object \n",
      " 17  dob                    1296675 non-null  object \n",
      " 18  trans_num              1296675 non-null  object \n",
      " 19  unix_time              1296675 non-null  int64  \n",
      " 20  merch_lat              1296675 non-null  float64\n",
      " 21  merch_long             1296675 non-null  float64\n",
      " 22  is_fraud               1296675 non-null  int64  \n",
      "dtypes: float64(5), int64(6), object(12)\n",
      "memory usage: 227.5+ MB\n"
     ]
    }
   ],
   "source": [
    "data_train.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that there are 22 features    \n",
    "\"is_fraud\" is the dependent variable "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.2 Shape of Data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1296675, 23) (555719, 23)\n"
     ]
    }
   ],
   "source": [
    "print(data_train.shape,data_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.3 Taking a fraction to run the model faster\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 23) (27786, 23)\n"
     ]
    }
   ],
   "source": [
    "# taking smaller sample to run the model faster\n",
    "\n",
    "df_train= data_train.sample(frac = 0.1,random_state=1)\n",
    "df_test= data_test.sample(frac = 0.05,random_state=1)\n",
    "print(df_train.shape,df_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>trans_date_trans_time</th>\n",
       "      <th>cc_num</th>\n",
       "      <th>merchant</th>\n",
       "      <th>category</th>\n",
       "      <th>amt</th>\n",
       "      <th>first</th>\n",
       "      <th>last</th>\n",
       "      <th>gender</th>\n",
       "      <th>street</th>\n",
       "      <th>...</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>job</th>\n",
       "      <th>dob</th>\n",
       "      <th>trans_num</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "      <th>is_fraud</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>94102</th>\n",
       "      <td>94102</td>\n",
       "      <td>2019-02-25 08:24:40</td>\n",
       "      <td>374497717543058</td>\n",
       "      <td>fraud_Funk Group</td>\n",
       "      <td>grocery_net</td>\n",
       "      <td>20.00</td>\n",
       "      <td>Linda</td>\n",
       "      <td>Hurst</td>\n",
       "      <td>F</td>\n",
       "      <td>31701 Tucker Square Suite 893</td>\n",
       "      <td>...</td>\n",
       "      <td>47.1709</td>\n",
       "      <td>-100.7944</td>\n",
       "      <td>1190</td>\n",
       "      <td>Designer, ceramics/pottery</td>\n",
       "      <td>1948-06-30</td>\n",
       "      <td>1595dec12f6f19ceaae9b7df0f8af5c0</td>\n",
       "      <td>1330158280</td>\n",
       "      <td>46.398331</td>\n",
       "      <td>-99.813959</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>198791</th>\n",
       "      <td>198791</td>\n",
       "      <td>2019-04-12 19:50:15</td>\n",
       "      <td>4428154703770710</td>\n",
       "      <td>fraud_Prosacco, Kreiger and Kovacek</td>\n",
       "      <td>home</td>\n",
       "      <td>284.88</td>\n",
       "      <td>Brittany</td>\n",
       "      <td>Guerra</td>\n",
       "      <td>F</td>\n",
       "      <td>79209 Gary Dale</td>\n",
       "      <td>...</td>\n",
       "      <td>46.5901</td>\n",
       "      <td>-117.1692</td>\n",
       "      <td>761</td>\n",
       "      <td>Chief Marketing Officer</td>\n",
       "      <td>1943-06-30</td>\n",
       "      <td>0ed26b649ed0fce94d8e632b7208dea0</td>\n",
       "      <td>1334260215</td>\n",
       "      <td>45.687331</td>\n",
       "      <td>-117.488135</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1238587</th>\n",
       "      <td>1238587</td>\n",
       "      <td>2020-05-31 21:50:53</td>\n",
       "      <td>213148039875802</td>\n",
       "      <td>fraud_Langworth, Boehm and Gulgowski</td>\n",
       "      <td>shopping_net</td>\n",
       "      <td>5.07</td>\n",
       "      <td>Jill</td>\n",
       "      <td>Jacobs</td>\n",
       "      <td>F</td>\n",
       "      <td>034 Kimberly Mountains</td>\n",
       "      <td>...</td>\n",
       "      <td>27.9551</td>\n",
       "      <td>-82.2966</td>\n",
       "      <td>79613</td>\n",
       "      <td>Environmental consultant</td>\n",
       "      <td>1978-11-30</td>\n",
       "      <td>7096316ec1a4b261e8613013827abae7</td>\n",
       "      <td>1370037053</td>\n",
       "      <td>27.254081</td>\n",
       "      <td>-81.974799</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Unnamed: 0 trans_date_trans_time            cc_num  \\\n",
       "94102         94102   2019-02-25 08:24:40   374497717543058   \n",
       "198791       198791   2019-04-12 19:50:15  4428154703770710   \n",
       "1238587     1238587   2020-05-31 21:50:53   213148039875802   \n",
       "\n",
       "                                     merchant      category     amt     first  \\\n",
       "94102                        fraud_Funk Group   grocery_net   20.00     Linda   \n",
       "198791    fraud_Prosacco, Kreiger and Kovacek          home  284.88  Brittany   \n",
       "1238587  fraud_Langworth, Boehm and Gulgowski  shopping_net    5.07      Jill   \n",
       "\n",
       "           last gender                         street  ...      lat      long  \\\n",
       "94102     Hurst      F  31701 Tucker Square Suite 893  ...  47.1709 -100.7944   \n",
       "198791   Guerra      F                79209 Gary Dale  ...  46.5901 -117.1692   \n",
       "1238587  Jacobs      F         034 Kimberly Mountains  ...  27.9551  -82.2966   \n",
       "\n",
       "         city_pop                         job         dob  \\\n",
       "94102        1190  Designer, ceramics/pottery  1948-06-30   \n",
       "198791        761     Chief Marketing Officer  1943-06-30   \n",
       "1238587     79613    Environmental consultant  1978-11-30   \n",
       "\n",
       "                                trans_num   unix_time  merch_lat  merch_long  \\\n",
       "94102    1595dec12f6f19ceaae9b7df0f8af5c0  1330158280  46.398331  -99.813959   \n",
       "198791   0ed26b649ed0fce94d8e632b7208dea0  1334260215  45.687331 -117.488135   \n",
       "1238587  7096316ec1a4b261e8613013827abae7  1370037053  27.254081  -81.974799   \n",
       "\n",
       "         is_fraud  \n",
       "94102           0  \n",
       "198791          0  \n",
       "1238587         0  \n",
       "\n",
       "[3 rows x 23 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.4 Checking the null values\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Unnamed: 0               0\n",
       "trans_date_trans_time    0\n",
       "cc_num                   0\n",
       "merchant                 0\n",
       "category                 0\n",
       "amt                      0\n",
       "first                    0\n",
       "last                     0\n",
       "gender                   0\n",
       "street                   0\n",
       "city                     0\n",
       "state                    0\n",
       "zip                      0\n",
       "lat                      0\n",
       "long                     0\n",
       "city_pop                 0\n",
       "job                      0\n",
       "dob                      0\n",
       "trans_num                0\n",
       "unix_time                0\n",
       "merch_lat                0\n",
       "merch_long               0\n",
       "is_fraud                 0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train.isnull().sum()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Unnamed: 0               0\n",
       "trans_date_trans_time    0\n",
       "cc_num                   0\n",
       "merchant                 0\n",
       "category                 0\n",
       "amt                      0\n",
       "first                    0\n",
       "last                     0\n",
       "gender                   0\n",
       "street                   0\n",
       "city                     0\n",
       "state                    0\n",
       "zip                      0\n",
       "lat                      0\n",
       "long                     0\n",
       "city_pop                 0\n",
       "job                      0\n",
       "dob                      0\n",
       "trans_num                0\n",
       "unix_time                0\n",
       "merch_lat                0\n",
       "merch_long               0\n",
       "is_fraud                 0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that no null values are present."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.5 Correlation Matrix\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#get correlations of each features in dataset\n",
    "corrmat = df_train.corr()\n",
    "top_corr_features = corrmat.index\n",
    "plt.figure(figsize=(15,15))\n",
    "\n",
    "#plot heat map\n",
    "g=sns.heatmap(df_train[top_corr_features].corr(),annot=True,cmap=\"RdYlGn\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.6 Histograms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#visual representation of the data using histograms \n",
    "df_train.hist(figsize = (15, 15))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.7 Plotting the classes "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Frequency')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting the classes\n",
    "plt.figure(figsize=(5,5))\n",
    "count_classes = pd.value_counts(df_train['is_fraud'], sort = True)\n",
    "count_classes.plot(kind = 'bar', rot=0)\n",
    "plt.title(\"Transaction Class Distribution\")\n",
    "LABELS = ['Normal','Fraud']\n",
    "plt.xticks(range(2), LABELS)\n",
    "plt.xlabel(\"Class of transaction\")\n",
    "plt.ylabel(\"Frequency\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.FacetGrid at 0x7ff128d69c50>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 407.75x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.relplot(y = 'unix_time', x = 'amt' , hue = 'is_fraud',data = df_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#sns.catplot(y = 'unix_time', x = 'amt' , hue = 'is_fraud',data = df_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that data is imbalanced    \n",
    "#### 2.8  Getting the Fraud and the Normal  transaction numbers for test and train dataset\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normal cases in train set : 128963 \n",
      "Fraud cases in train set : 705\n",
      "Normal cases in test set : 27656 \n",
      "Fraud cases in test set : 130\n"
     ]
    }
   ],
   "source": [
    "## Get the Fraud and the normal  transaction numbers for test and train dataset \n",
    "\n",
    "fraud_train = df_train[df_train['is_fraud']==1]\n",
    "normal_train = df_train[df_train['is_fraud']==0]\n",
    "fraud_test = df_test[df_test['is_fraud']==1]\n",
    "normal_test = df_test[df_test['is_fraud']==0]\n",
    "\n",
    "print(\"Normal cases in train set :\",len(df_train)-len(fraud_train),\"\\nFraud cases in train set :\",len(fraud_train))\n",
    "print(\"Normal cases in test set :\",len(df_test)-len(fraud_test),\"\\nFraud cases in test set :\",len(fraud_test))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "## 3. Data transformation and feature engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "From the exploratory data Analysis we make the following observations:   \n",
    "\n",
    "1. The first column contains just the indices and is not useful so we will drop it.   \n",
    "2. The third column with customer card number is also not useful , so we will drop it. \n",
    "3. \"first name\" and \"last name\" can also be dropped.  \n",
    "4. Transaction number - is it really needed? can be dropped.\n",
    "5. Date time column can be used to calculate the age of the customer\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.1 Dropping the columns not needed\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function to drop tbe columns\n",
    "def dropCol(data):\n",
    "    col_to_drop = ['trans_date_trans_time','Unnamed: 0','cc_num','first','last','trans_num']\n",
    "    res = data.drop(col_to_drop,axis = 1)\n",
    "    return res\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 17) (27786, 17)\n"
     ]
    }
   ],
   "source": [
    "# dropping the columns\n",
    "# dropping the columns ['trans_date_trans_time','Unnamed: 0','cc_num','first','last','trans_num']\n",
    "# train data set\n",
    "df_train = dropCol(df_train)\n",
    "# test data set\n",
    "df_test = dropCol(df_test)\n",
    "\n",
    "print ( df_train.shape, df_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.2 Dropping the columns not needed\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 16) (129668,) (27786, 16) (27786,)\n"
     ]
    }
   ],
   "source": [
    "#Create independent and Dependent Features\n",
    "columns = df_train.columns.tolist()\n",
    "\n",
    "# removing the dependent feature is_fraud\n",
    "columns = [c for c in columns if c not in [\"is_fraud\"]]\n",
    "\n",
    "X_train = df_train[columns]\n",
    "Y_train = df_train['is_fraud']\n",
    "X_test = df_test[columns]\n",
    "Y_test = df_test['is_fraud']\n",
    "print ( X_train.shape, Y_train.shape,X_test.shape, Y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.3 Converting dob to age\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 16) (27786, 16)\n"
     ]
    }
   ],
   "source": [
    "# function to convert dob to years\n",
    "def age_years(born):\n",
    "    return 2019 - int(born[0:4])\n",
    "\n",
    "# replacing the dob column with age column in our data set for test and train\n",
    "X_train['age'] = X_train['dob'].apply(lambda x: age_years(x))\n",
    "X_train = X_train.drop(['dob'],axis =1)\n",
    "\n",
    "X_test['age'] = X_test['dob'].apply(lambda x: age_years(x))\n",
    "X_test = X_test.drop(['dob'],axis =1)\n",
    "print(X_train.shape,X_test.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>merchant</th>\n",
       "      <th>category</th>\n",
       "      <th>amt</th>\n",
       "      <th>gender</th>\n",
       "      <th>street</th>\n",
       "      <th>city</th>\n",
       "      <th>state</th>\n",
       "      <th>zip</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>job</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>94102</th>\n",
       "      <td>fraud_Funk Group</td>\n",
       "      <td>grocery_net</td>\n",
       "      <td>20.00</td>\n",
       "      <td>F</td>\n",
       "      <td>31701 Tucker Square Suite 893</td>\n",
       "      <td>Wilton</td>\n",
       "      <td>ND</td>\n",
       "      <td>58579</td>\n",
       "      <td>47.1709</td>\n",
       "      <td>-100.7944</td>\n",
       "      <td>1190</td>\n",
       "      <td>Designer, ceramics/pottery</td>\n",
       "      <td>1330158280</td>\n",
       "      <td>46.398331</td>\n",
       "      <td>-99.813959</td>\n",
       "      <td>71</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>198791</th>\n",
       "      <td>fraud_Prosacco, Kreiger and Kovacek</td>\n",
       "      <td>home</td>\n",
       "      <td>284.88</td>\n",
       "      <td>F</td>\n",
       "      <td>79209 Gary Dale</td>\n",
       "      <td>Colton</td>\n",
       "      <td>WA</td>\n",
       "      <td>99113</td>\n",
       "      <td>46.5901</td>\n",
       "      <td>-117.1692</td>\n",
       "      <td>761</td>\n",
       "      <td>Chief Marketing Officer</td>\n",
       "      <td>1334260215</td>\n",
       "      <td>45.687331</td>\n",
       "      <td>-117.488135</td>\n",
       "      <td>76</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1238587</th>\n",
       "      <td>fraud_Langworth, Boehm and Gulgowski</td>\n",
       "      <td>shopping_net</td>\n",
       "      <td>5.07</td>\n",
       "      <td>F</td>\n",
       "      <td>034 Kimberly Mountains</td>\n",
       "      <td>Brandon</td>\n",
       "      <td>FL</td>\n",
       "      <td>33510</td>\n",
       "      <td>27.9551</td>\n",
       "      <td>-82.2966</td>\n",
       "      <td>79613</td>\n",
       "      <td>Environmental consultant</td>\n",
       "      <td>1370037053</td>\n",
       "      <td>27.254081</td>\n",
       "      <td>-81.974799</td>\n",
       "      <td>41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>619078</th>\n",
       "      <td>fraud_Conroy-Emard</td>\n",
       "      <td>food_dining</td>\n",
       "      <td>45.38</td>\n",
       "      <td>F</td>\n",
       "      <td>9345 Spencer Junctions Suite 183</td>\n",
       "      <td>Alpharetta</td>\n",
       "      <td>GA</td>\n",
       "      <td>30009</td>\n",
       "      <td>34.0770</td>\n",
       "      <td>-84.3033</td>\n",
       "      <td>165556</td>\n",
       "      <td>Prison officer</td>\n",
       "      <td>1348160812</td>\n",
       "      <td>34.551957</td>\n",
       "      <td>-83.374265</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>573850</th>\n",
       "      <td>fraud_Adams-Barrows</td>\n",
       "      <td>health_fitness</td>\n",
       "      <td>27.52</td>\n",
       "      <td>M</td>\n",
       "      <td>09419 Lewis Rapids</td>\n",
       "      <td>Bay City</td>\n",
       "      <td>MI</td>\n",
       "      <td>48706</td>\n",
       "      <td>43.6088</td>\n",
       "      <td>-83.9530</td>\n",
       "      <td>67858</td>\n",
       "      <td>Water quality scientist</td>\n",
       "      <td>1346455370</td>\n",
       "      <td>43.032957</td>\n",
       "      <td>-83.521294</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>944454</th>\n",
       "      <td>fraud_Reichert-Weissnat</td>\n",
       "      <td>travel</td>\n",
       "      <td>5.83</td>\n",
       "      <td>F</td>\n",
       "      <td>6911 Nicholas Keys Apt. 237</td>\n",
       "      <td>Armagh</td>\n",
       "      <td>PA</td>\n",
       "      <td>15920</td>\n",
       "      <td>40.4603</td>\n",
       "      <td>-79.0097</td>\n",
       "      <td>922</td>\n",
       "      <td>Early years teacher</td>\n",
       "      <td>1358007774</td>\n",
       "      <td>40.522369</td>\n",
       "      <td>-79.314668</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23683</th>\n",
       "      <td>fraud_Baumbach, Strosin and Nicolas</td>\n",
       "      <td>shopping_pos</td>\n",
       "      <td>1.96</td>\n",
       "      <td>F</td>\n",
       "      <td>27479 Reeves Dale</td>\n",
       "      <td>Whaleyville</td>\n",
       "      <td>MD</td>\n",
       "      <td>21872</td>\n",
       "      <td>38.4121</td>\n",
       "      <td>-75.2811</td>\n",
       "      <td>718</td>\n",
       "      <td>Regulatory affairs officer</td>\n",
       "      <td>1326572609</td>\n",
       "      <td>38.534192</td>\n",
       "      <td>-76.161654</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296375</th>\n",
       "      <td>fraud_Lubowitz-Walter</td>\n",
       "      <td>kids_pets</td>\n",
       "      <td>26.59</td>\n",
       "      <td>F</td>\n",
       "      <td>72966 Shannon Pass Apt. 391</td>\n",
       "      <td>Bauxite</td>\n",
       "      <td>AR</td>\n",
       "      <td>72011</td>\n",
       "      <td>34.5091</td>\n",
       "      <td>-92.4828</td>\n",
       "      <td>4074</td>\n",
       "      <td>Scientist, audiological</td>\n",
       "      <td>1337963080</td>\n",
       "      <td>35.175647</td>\n",
       "      <td>-91.740062</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>306312</th>\n",
       "      <td>fraud_Runolfsdottir, Mueller and Hand</td>\n",
       "      <td>entertainment</td>\n",
       "      <td>6.00</td>\n",
       "      <td>M</td>\n",
       "      <td>3283 James Station</td>\n",
       "      <td>Fayetteville</td>\n",
       "      <td>NC</td>\n",
       "      <td>28314</td>\n",
       "      <td>35.0583</td>\n",
       "      <td>-79.0080</td>\n",
       "      <td>238602</td>\n",
       "      <td>Scientist, research (maths)</td>\n",
       "      <td>1338216727</td>\n",
       "      <td>34.360264</td>\n",
       "      <td>-78.128199</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1171139</th>\n",
       "      <td>fraud_Boehm, Predovic and Reinger</td>\n",
       "      <td>misc_pos</td>\n",
       "      <td>44.70</td>\n",
       "      <td>M</td>\n",
       "      <td>8415 Vaughn Squares Apt. 788</td>\n",
       "      <td>Acworth</td>\n",
       "      <td>NH</td>\n",
       "      <td>3601</td>\n",
       "      <td>43.1960</td>\n",
       "      <td>-72.3001</td>\n",
       "      <td>477</td>\n",
       "      <td>Naval architect</td>\n",
       "      <td>1367637527</td>\n",
       "      <td>44.156195</td>\n",
       "      <td>-73.285582</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>129668 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                      merchant        category     amt gender  \\\n",
       "94102                         fraud_Funk Group     grocery_net   20.00      F   \n",
       "198791     fraud_Prosacco, Kreiger and Kovacek            home  284.88      F   \n",
       "1238587   fraud_Langworth, Boehm and Gulgowski    shopping_net    5.07      F   \n",
       "619078                      fraud_Conroy-Emard     food_dining   45.38      F   \n",
       "573850                     fraud_Adams-Barrows  health_fitness   27.52      M   \n",
       "...                                        ...             ...     ...    ...   \n",
       "944454                 fraud_Reichert-Weissnat          travel    5.83      F   \n",
       "23683      fraud_Baumbach, Strosin and Nicolas    shopping_pos    1.96      F   \n",
       "296375                   fraud_Lubowitz-Walter       kids_pets   26.59      F   \n",
       "306312   fraud_Runolfsdottir, Mueller and Hand   entertainment    6.00      M   \n",
       "1171139      fraud_Boehm, Predovic and Reinger        misc_pos   44.70      M   \n",
       "\n",
       "                                   street          city state    zip      lat  \\\n",
       "94102       31701 Tucker Square Suite 893        Wilton    ND  58579  47.1709   \n",
       "198791                    79209 Gary Dale        Colton    WA  99113  46.5901   \n",
       "1238587            034 Kimberly Mountains       Brandon    FL  33510  27.9551   \n",
       "619078   9345 Spencer Junctions Suite 183    Alpharetta    GA  30009  34.0770   \n",
       "573850                 09419 Lewis Rapids      Bay City    MI  48706  43.6088   \n",
       "...                                   ...           ...   ...    ...      ...   \n",
       "944454        6911 Nicholas Keys Apt. 237        Armagh    PA  15920  40.4603   \n",
       "23683                   27479 Reeves Dale   Whaleyville    MD  21872  38.4121   \n",
       "296375        72966 Shannon Pass Apt. 391       Bauxite    AR  72011  34.5091   \n",
       "306312                 3283 James Station  Fayetteville    NC  28314  35.0583   \n",
       "1171139      8415 Vaughn Squares Apt. 788       Acworth    NH   3601  43.1960   \n",
       "\n",
       "             long  city_pop                          job   unix_time  \\\n",
       "94102   -100.7944      1190   Designer, ceramics/pottery  1330158280   \n",
       "198791  -117.1692       761      Chief Marketing Officer  1334260215   \n",
       "1238587  -82.2966     79613     Environmental consultant  1370037053   \n",
       "619078   -84.3033    165556               Prison officer  1348160812   \n",
       "573850   -83.9530     67858      Water quality scientist  1346455370   \n",
       "...           ...       ...                          ...         ...   \n",
       "944454   -79.0097       922          Early years teacher  1358007774   \n",
       "23683    -75.2811       718   Regulatory affairs officer  1326572609   \n",
       "296375   -92.4828      4074      Scientist, audiological  1337963080   \n",
       "306312   -79.0080    238602  Scientist, research (maths)  1338216727   \n",
       "1171139  -72.3001       477              Naval architect  1367637527   \n",
       "\n",
       "         merch_lat  merch_long  age  \n",
       "94102    46.398331  -99.813959   71  \n",
       "198791   45.687331 -117.488135   76  \n",
       "1238587  27.254081  -81.974799   41  \n",
       "619078   34.551957  -83.374265   22  \n",
       "573850   43.032957  -83.521294   24  \n",
       "...            ...         ...  ...  \n",
       "944454   40.522369  -79.314668   47  \n",
       "23683    38.534192  -76.161654   34  \n",
       "296375   35.175647  -91.740062   44  \n",
       "306312   34.360264  -78.128199   26  \n",
       "1171139  44.156195  -73.285582   31  \n",
       "\n",
       "[129668 rows x 16 columns]"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.4 Converting the categorical features to numerical by one- hot - encoding\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(157454, 16)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# concanating the test and train data so that number of columns remain the same in both the data sets\n",
    "final_df = pd.concat([X_train,X_test],axis=0)\n",
    "final_df.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['merchant', 'category', 'gender', 'street', 'city', 'state', 'job']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# creating the list of categorical variables\n",
    "categorical_features =[feature for feature in X_train.columns if final_df[feature].dtypes == 'O']\n",
    "categorical_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Distinct categories for merchant  are 693\n",
      "Distinct categories for category  are 14\n",
      "Distinct categories for gender  are 2\n",
      "Distinct categories for street  are 957\n",
      "Distinct categories for city  are 872\n",
      "Distinct categories for state  are 51\n",
      "Distinct categories for job  are 486\n"
     ]
    }
   ],
   "source": [
    "#observing the unique values in each feature\n",
    "for feature in categorical_features:\n",
    "    print(\"Distinct categories for {}  are {}\".format(feature,len(final_df[feature].unique())))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function to convert categorical variables to one hot encoding\n",
    "def category_onehot_multcols(data,multcolumns):\n",
    "    df_final = data\n",
    "    i=0\n",
    "    for fields in multcolumns:\n",
    "        print(fields)\n",
    "        df1=pd.get_dummies(final_df[fields],drop_first=True)\n",
    "        final_df.drop([fields],axis=1,inplace=True)\n",
    "        if i==0:\n",
    "            df_final=df1.copy()\n",
    "        else:           \n",
    "            df_final=pd.concat([df_final,df1],axis=1)\n",
    "        i=i+1             \n",
    "    df_final=pd.concat([final_df,df_final],axis=1)\n",
    "    return df_final"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "merchant\n",
      "category\n",
      "gender\n",
      "street\n",
      "city\n",
      "state\n",
      "job\n"
     ]
    }
   ],
   "source": [
    "# applying the one hot encoding\n",
    "final_df = category_onehot_multcols(final_df, categorical_features)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(157454, 3077)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# removing duplicated columns\n",
    "final_df =final_df.loc[:,~final_df.columns.duplicated()]\n",
    "final_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 3077) (27786, 3077)\n"
     ]
    }
   ],
   "source": [
    "# separating the test and training data\n",
    "df_Train=final_df.iloc[:129668,:]\n",
    "df_Test=final_df.iloc[129668:,:]\n",
    "print(df_Train.shape,df_Test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0    128963\n",
      "1       705\n",
      "Name: is_fraud, dtype: int64 \n",
      " 0    27656\n",
      "1      130\n",
      "Name: is_fraud, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(Y_train.value_counts(),\"\\n\",Y_test.value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 3077) (27786, 3077) (129668,) (27786,)\n"
     ]
    }
   ],
   "source": [
    "# files ready for testing on models\n",
    "print(df_Train.shape, df_Test.shape, Y_train.shape, Y_test.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.005466684242767306"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "705/128963"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we see that the data is skewed, i.e the number of samples for class 1 is less tha 0.5% of the samples of class 0.\n",
    "In this case the the machine learning algorithms with not predict the Fraud cases correctly.\n",
    "To predict the Fraud correctly we have two approaches:   \n",
    "1. Handle the imbalance in Data and apply various Machine Learning algorithms and then apply ML models.    \n",
    "2. Predict Fraud as outlier/Anomaly    \n",
    "3. Change detection is similar to anomaly detection except rather than looking for an anomaly it is looking for a change or difference. This might be a change in behavior of a user as observed by usage patterns or bank transactions.\n",
    "\n",
    "We proceed for the first method first.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Handling the Imbalance in data \n",
    "\n",
    "   \n",
    "Handling the Imbalance in data by one of the following methods:   \n",
    "> 4.1 Random Under-Sampling   \n",
    "       4.2 Random Over-Sampling   \n",
    "       4.3 SMOTE (Synthetic Minority over sampling technique)   \n",
    "       4.4 Near Miss algorighm ( under sampling )   \n",
    "       4.5 Ensemble method \n",
    "   \n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First we apply the model without handling the imbalance. We will use Logistic Regression and Decision Tree classifier for our exploration of the method to apply to handle imbalance in data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Logistic Regression\n",
    "model_LR = LogisticRegression(random_state=137)\n",
    "model_LR.fit(df_Train,Y_train)\n",
    "y_pred = model_LR.predict(df_Test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's evaluate our model \n",
    "def print_eval(y_pred,model):\n",
    "    print(\"Training Accuracy: \",model.score(df_Train, Y_train))\n",
    "    print(\"Testing Accuracy: \", model.score(df_Test, Y_test))\n",
    "    cm = confusion_matrix(Y_test, y_pred)\n",
    "    print(cm)\n",
    "    print(classification_report(Y_test,y_pred))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[27656     0]\n",
      " [  130     0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.00      0.00      0.00       130\n",
      "\n",
      "    accuracy                           1.00     27786\n",
      "   macro avg       0.50      0.50      0.50     27786\n",
      "weighted avg       0.99      1.00      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Logistic Regression Classification results without balancong class\n",
    "print_eval(y_pred,model_LR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  1.0\n",
      "Testing Accuracy:  0.9866839415533002\n",
      "[[27333   323]\n",
      " [   47    83]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.99      0.99     27656\n",
      "           1       0.20      0.64      0.31       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.60      0.81      0.65     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# decision tree\n",
    "decision_tree_model = DecisionTreeClassifier(random_state=137)\n",
    "decision_tree_model.fit(df_Train,Y_train)\n",
    "y_pred = decision_tree_model.predict(df_Test)\n",
    "# Decision Tree Classification results without balancong class\n",
    "print_eval(y_pred,decision_tree_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n"
     ]
    }
   ],
   "source": [
    "# helper function to print a comparitive results of all models\n",
    "model_list = {}\n",
    "def getComparitiveChart(model_list,model_name,model):\n",
    "    print('\\033[1m', \" Comparitive Chart of all models\")\n",
    "    if model_name not in model_list:\n",
    "        model_list[model_name]= model\n",
    "    \n",
    "    print( \" Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\")\n",
    "    for i,v in model_list.items():\n",
    "        pred = v.predict(df_Test)\n",
    "        prfs = precision_recall_fscore_support(Y_test,pred , average=None)\n",
    "        #res_recall[i] = round(prfs[1][1],2)\n",
    "        recall_0   = prfs[1][0]\n",
    "        recall_1   = prfs[1][1]\n",
    "        f1_score_0 = prfs[2][0]\n",
    "        f1_score_1 = prfs[2][1]\n",
    "\n",
    "        print(f'{f\"{i}\":<35}' ,\"{:.2f}\".format(recall_1),\"\\t\\t\",\"{:.2f}\".format(recall_0),\n",
    "               \"\\t\\t\",\"{:.2f}\".format(f1_score_1),\"\\t\\t\",\"{:.2f}\".format(f1_score_0) )\n",
    "        \n",
    "getComparitiveChart(model_list,'Logistic Regression',model_LR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Decision Tree',decision_tree_model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Metric Analysis\n",
    "\n",
    "1. **True  Positives** : Correctly classified as Safe Transaction     = 27333 (.99)   \n",
    "2. **False Negitives** : Mis-classified Safe Transaction              = 323 (Harmless)   \n",
    "3. **False Positives** : Mis-classified as Fraud Transactions         = 47 (Dangerous )   \n",
    "4. **True Negatives**  : Correctly classified as Fraud Transactions   = 83 (out of 130 -> .64)   \n",
    "5. **Accuracy**        : .99   \n",
    "\n",
    "Here the accuracy is not taken into account as it is misleading.   \n",
    "We want to get maximum True Negatives i.e we want to predict the Fraud tranactions with maximum accuracy. This can be done by monitoring the Recall.\n",
    "\n",
    "So when the data is imbalanced the Recall is    \n",
    "class 0 (Safe) : 0.99   \n",
    "class 1 (Fraud): 0.64   \n",
    "\n",
    "**We want the recall of class 1 to be close to 1.00**    \n",
    "**We can also use f1-score as a metric to observe**\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.1 Implementing Random Under-sampling\n",
    "\n",
    "**Advantages**   \n",
    "- It can help improve run time and storage problems by reducing the number of training data samples when the training data set is huge.   \n",
    "\n",
    "**Disadvantages**   \n",
    "- It can discard potentially useful information which could be important for building rule classifiers.\n",
    "The sample chosen by random under sampling may be a biased sample. And it will not be an accurate representative of the population. Thereby, resulting in inaccurate results with the actual test data set.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668,) (129668, 3077)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(129668, 3078)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# adding the dependent feature in the train data set\n",
    "print(Y_train.shape,df_Train.shape)\n",
    "df_train = pd.concat([df_Train,Y_train],axis = 1)\n",
    "df_train.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "128963 705\n",
      "(128963, 3078) (705, 3078)\n"
     ]
    }
   ],
   "source": [
    "# Class count\n",
    "count_class_0, count_class_1 = df_train.is_fraud.value_counts()\n",
    "print(count_class_0, count_class_1)\n",
    "# Divide by class\n",
    "df_class_0 = df_train[df_train['is_fraud'] == 0]\n",
    "df_class_1 = df_train[df_train['is_fraud'] == 1]\n",
    "print(df_class_0.shape,df_class_1.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random under-sampling:\n",
      "1    705\n",
      "0    705\n",
      "Name: is_fraud, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# Undersample 0-class and concat the DataFrames of both class\n",
    "df_class_0_under = df_class_0.sample(count_class_1)\n",
    "df_train_under_sample = pd.concat([df_class_0_under, df_class_1], axis=0)\n",
    "\n",
    "print('Random under-sampling:')\n",
    "print(df_train_under_sample.is_fraud.value_counts())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.843909060061079\n",
      "Testing Accuracy:  0.8447419563809112\n",
      "[[23455  4201]\n",
      " [  113    17]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.85      0.92     27656\n",
      "           1       0.00      0.13      0.01       130\n",
      "\n",
      "    accuracy                           0.84     27786\n",
      "   macro avg       0.50      0.49      0.46     27786\n",
      "weighted avg       0.99      0.84      0.91     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : Logistic Regression Under Sampling \n",
    "X = df_train_under_sample.drop('is_fraud',axis='columns')\n",
    "y = df_train_under_sample['is_fraud']\n",
    "model_LR_under_sample = LogisticRegression(random_state=137)\n",
    "model_LR_under_sample.fit(X,y)\n",
    "y_pred = model_LR_under_sample.predict(df_Test)\n",
    "print_eval(y_pred,model_LR_under_sample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9302140852022087\n",
      "Testing Accuracy:  0.9329878356006622\n",
      "[[25797  1859]\n",
      " [    3   127]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.93      0.97     27656\n",
      "           1       0.06      0.98      0.12       130\n",
      "\n",
      "    accuracy                           0.93     27786\n",
      "   macro avg       0.53      0.95      0.54     27786\n",
      "weighted avg       1.00      0.93      0.96     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : decision tree Regression Under Sampling\n",
    "decision_tree_model_undersample = DecisionTreeClassifier(random_state=137)\n",
    "decision_tree_model_undersample.fit(X,y)\n",
    "y_pred = decision_tree_model_undersample.predict(df_Test)\n",
    "print_eval(y_pred,decision_tree_model_undersample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Logistic Regression Under Sample',model_LR_under_sample)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n",
      "Decision Tree Under Sample          0.98 \t\t 0.93 \t\t 0.12 \t\t 0.97\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Decision Tree Under Sample',decision_tree_model_undersample)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Comparision of results and Observations :\n",
    "\n",
    "|                                       | Class-0 Recall| Class-1 Recall|Class-0 f1-score| Class-1 f1-score|\n",
    "|---------------------------------------|---------------|---------------|----------------|-----------------|\n",
    "|Logistic Regression                    |     1.00      |     0.00      |     1.00       |      0.00       |\n",
    "|Decision Tree Classifier               |     0.99      |     0.64      |     0.99       |      0.31       |\n",
    "|Logistic Regression Under Sampling     |     1.00      |     0.00      |     1.00       |      0.50       |\n",
    "|Decision Tree Classifier Under Sampling|     0.94      |     0.95      |     0.97       |      0.13       |\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.2 Implementing Random Over-sampling\n",
    "\n",
    "**Advantages**\n",
    "- Unlike under sampling this method leads to no information loss.\n",
    "- Outperforms under sampling  \n",
    "\n",
    "**Disadvantages**\n",
    "- It increases the likelihood of overfitting since it replicates the minority class events.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "128963 705\n",
      "(128963, 3078) (705, 3078)\n"
     ]
    }
   ],
   "source": [
    "# Class count\n",
    "count_class_0, count_class_1 = df_train.is_fraud.value_counts()\n",
    "print(count_class_0, count_class_1)\n",
    "# Divide by class\n",
    "df_class_0 = df_train[df_train['is_fraud'] == 0]\n",
    "df_class_1 = df_train[df_train['is_fraud'] == 1]\n",
    "print(df_class_0.shape,df_class_1.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random over-sampling:\n",
      "1    128963\n",
      "0    128963\n",
      "Name: is_fraud, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# Oversample 1-class and concat the DataFrames of both class\n",
    "df_class_1_over = df_class_1.sample(count_class_0,replace=True)\n",
    "df_train_over_sample = pd.concat([df_class_0, df_class_1_over], axis=0)\n",
    "\n",
    "print('Random over-sampling:')\n",
    "\n",
    "print(df_train_over_sample.is_fraud.value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.8594564580312799\n",
      "Testing Accuracy:  0.8624847045274598\n",
      "[[23953  3703]\n",
      " [  118    12]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.87      0.93     27656\n",
      "           1       0.00      0.09      0.01       130\n",
      "\n",
      "    accuracy                           0.86     27786\n",
      "   macro avg       0.50      0.48      0.47     27786\n",
      "weighted avg       0.99      0.86      0.92     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : Logistic Regression\n",
    "X = df_train_over_sample.drop('is_fraud',axis='columns')\n",
    "y = df_train_over_sample['is_fraud']\n",
    "model_LR_over_sample = LogisticRegression(random_state=137)\n",
    "model_LR_over_sample.fit(X,y)\n",
    "y_pred = model_LR_over_sample.predict(df_Test)\n",
    "print_eval(y_pred,model_LR_over_sample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  1.0\n",
      "Testing Accuracy:  0.9938818109839488\n",
      "[[27583    73]\n",
      " [   97    33]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.31      0.25      0.28       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.65      0.63      0.64     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : decision tree\n",
    "decision_tree_model_over_sample = DecisionTreeClassifier(random_state=137)\n",
    "decision_tree_model_over_sample.fit(X,y)\n",
    "y_pred = decision_tree_model_over_sample.predict(df_Test)\n",
    "print_eval(y_pred,decision_tree_model_over_sample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n",
      "Decision Tree Under Sample          0.98 \t\t 0.93 \t\t 0.12 \t\t 0.97\n",
      "Logistic Regression Over Sample     0.09 \t\t 0.87 \t\t 0.01 \t\t 0.93\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Logistic Regression Over Sample',model_LR_over_sample)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n",
      "Decision Tree Under Sample          0.98 \t\t 0.93 \t\t 0.12 \t\t 0.97\n",
      "Logistic Regression Over Sample     0.09 \t\t 0.87 \t\t 0.01 \t\t 0.93\n",
      "Decision Tree Over Sample           0.25 \t\t 1.00 \t\t 0.28 \t\t 1.00\n"
     ]
    }
   ],
   "source": [
    "\n",
    "getComparitiveChart(model_list,'Decision Tree Over Sample',decision_tree_model_over_sample)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.3 Implementing SMOTE (Synthetic Minority Over Sampling Technique) - Informed Over Sampling\n",
    "\n",
    "SMOTE algorithm works in 4 simple steps:\n",
    "\n",
    "- Choose a minority class as the input vector\n",
    "- Find its k nearest neighbors (k_neighbors is specified as an argument in the SMOTE() function,defalut value is 5)\n",
    "- Choose one of these neighbors and place a synthetic point anywhere on the line joining the point under consideration and its chosen neighbor\n",
    "- Repeat the steps until data is balanced\n",
    "\n",
    "**Advantages**\n",
    "- Mitigates the problem of overfitting caused by random oversampling as synthetic examples are generated rather than replication of instances   \n",
    "- No loss of useful information   \n",
    "\n",
    "**Disadvantages**\n",
    "- While generating synthetic examples SMOTE does not take into consideration neighboring examples from other classes. This can result in increase in overlapping of classes and can introduce additional noise   \n",
    "- SMOTE is not very effective for high dimensional data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 3077) (27786, 3077)\n"
     ]
    }
   ],
   "source": [
    "print(df_Train.shape,df_Test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    128963\n",
       "0    128963\n",
       "Name: is_fraud, dtype: int64"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# smote implementation\n",
    "smote = SMOTE(sampling_strategy = 'minority')\n",
    "X_sm, y_sm = smote.fit_sample(df_Train, Y_train)\n",
    "y_sm.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.4928972452725422\n",
      "Testing Accuracy:  0.47970200820557113\n",
      "[[13272 14384]\n",
      " [   73    57]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.99      0.48      0.65     27656\n",
      "           1       0.00      0.44      0.01       130\n",
      "\n",
      "    accuracy                           0.48     27786\n",
      "   macro avg       0.50      0.46      0.33     27786\n",
      "weighted avg       0.99      0.48      0.64     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : Logistic Regression\n",
    "\n",
    "model_LR_smote = LogisticRegression(random_state=137)\n",
    "model_LR_smote.fit(X_sm,y_sm)\n",
    "y_pred = model_LR_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_LR_smote)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  1.0\n",
      "Testing Accuracy:  0.991542503418988\n",
      "[[27478   178]\n",
      " [   57    73]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.99      1.00     27656\n",
      "           1       0.29      0.56      0.38       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.64      0.78      0.69     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : decision tree\n",
    "decision_tree_model_smote = DecisionTreeClassifier(random_state=137)\n",
    "decision_tree_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = decision_tree_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,decision_tree_model_smote)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n",
      "Decision Tree Under Sample          0.98 \t\t 0.93 \t\t 0.12 \t\t 0.97\n",
      "Logistic Regression Over Sample     0.09 \t\t 0.87 \t\t 0.01 \t\t 0.93\n",
      "Decision Tree Over Sample           0.25 \t\t 1.00 \t\t 0.28 \t\t 1.00\n",
      "Logistic Regression SMOTE           0.44 \t\t 0.48 \t\t 0.01 \t\t 0.65\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Logistic Regression SMOTE',model_LR_smote)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n",
      "Decision Tree Under Sample          0.98 \t\t 0.93 \t\t 0.12 \t\t 0.97\n",
      "Logistic Regression Over Sample     0.09 \t\t 0.87 \t\t 0.01 \t\t 0.93\n",
      "Decision Tree Over Sample           0.25 \t\t 1.00 \t\t 0.28 \t\t 1.00\n",
      "Logistic Regression SMOTE           0.44 \t\t 0.48 \t\t 0.01 \t\t 0.65\n",
      "Decision Tree SMOTE                 0.56 \t\t 0.99 \t\t 0.38 \t\t 1.00\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Decision Tree SMOTE',decision_tree_model_smote)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "\n",
    "**We see that Under sampling gives better results till now**\n",
    "### 4.4 Implementing Near Miss\n",
    "\n",
    "NearMiss is an under-sampling technique.    \n",
    "Instead of resampling the Minority class, using a distance, this will make the majority class equal to the minority class.When instances of two different classes are very close to each other, we remove the instances of the majority class to increase the spaces between the two classes. This helps in the classification process.   \n",
    "Steps in Near Miss Algorithm.  \n",
    "- Step 1: The method first finds the distances between all instances of the majority class and the instances of the minority class. Here, majority class is to be under-sampled.\n",
    "- Step 2: Then, n instances of the majority class that have the smallest distances to those in the minority class are selected.\n",
    "- Step 3: If there are k instances in the minority class, the nearest method will result in k*n instances of the majority class.\n",
    "\n",
    "\n",
    "For finding n closest instances in the majority class, there are several variations of applying NearMiss Algorithm :   \n",
    "\n",
    "- NearMiss – Version 1 : It selects samples of the majority class for which average distances to the k closest instances of the minority class is smallest.\n",
    "- NearMiss – Version 2 : It selects samples of the majority class for which average distances to the k farthest instances of the minority class is smallest.\n",
    "- NearMiss – Version 3 : It works in 2 steps. Firstly, for each minority class instance, their M nearest-neighbors will be stored. Then finally, the majority class instances are selected for which the average distance to the N nearest-neighbors is the largest."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Near Miss:\n",
      "1    705\n",
      "0    705\n",
      "Name: is_fraud, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# near miss\n",
    "nr = NearMiss() \n",
    "X_train_miss, y_train_miss = nr.fit_sample(df_Train, Y_train) \n",
    "print('Near Miss:')\n",
    "print(y_train_miss.value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    705\n",
      "0    705\n",
      "Name: is_fraud, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(y_train_miss.value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.6187185735879323\n",
      "Testing Accuracy:  0.6264665658964946\n",
      "[[17361 10295]\n",
      " [   84    46]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.63      0.77     27656\n",
      "           1       0.00      0.35      0.01       130\n",
      "\n",
      "    accuracy                           0.63     27786\n",
      "   macro avg       0.50      0.49      0.39     27786\n",
      "weighted avg       0.99      0.63      0.77     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : Logistic Regression\n",
    "\n",
    "model_LR_nm = LogisticRegression(random_state=137)\n",
    "model_LR_nm.fit(X_train_miss,y_train_miss)\n",
    "y_pred = model_LR_nm.predict(df_Test)\n",
    "print_eval(y_pred,model_LR_nm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### trade off between precision and recall\n",
    "\n",
    "https://www.kaggle.com/robinteuwens/precision-vs-recall-optimizing-fraud-costs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.6950596909029213\n",
      "Testing Accuracy:  0.696681782192471\n",
      "[[19236  8420]\n",
      " [    8   122]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.70      0.82     27656\n",
      "           1       0.01      0.94      0.03       130\n",
      "\n",
      "    accuracy                           0.70     27786\n",
      "   macro avg       0.51      0.82      0.42     27786\n",
      "weighted avg       0.99      0.70      0.82     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# training and predictions : decision tree\n",
    "decision_tree_model_nm = DecisionTreeClassifier(random_state=137)\n",
    "decision_tree_model_nm.fit(X_train_miss,y_train_miss)\n",
    "y_pred = decision_tree_model_nm.predict(df_Test)\n",
    "print_eval(y_pred,decision_tree_model_nm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n",
      "Decision Tree Under Sample          0.98 \t\t 0.93 \t\t 0.12 \t\t 0.97\n",
      "Logistic Regression Over Sample     0.09 \t\t 0.87 \t\t 0.01 \t\t 0.93\n",
      "Decision Tree Over Sample           0.25 \t\t 1.00 \t\t 0.28 \t\t 1.00\n",
      "Logistic Regression SMOTE           0.44 \t\t 0.48 \t\t 0.01 \t\t 0.65\n",
      "Decision Tree SMOTE                 0.56 \t\t 0.99 \t\t 0.38 \t\t 1.00\n",
      "Logistic Regression Near Miss       0.35 \t\t 0.63 \t\t 0.01 \t\t 0.77\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Logistic Regression Near Miss',model_LR_nm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m  Comparitive Chart of all models\n",
      " Model                             Recall class-1 Recall class-0 f1-score class-1 f1-score class-0\n",
      "Logistic Regression                 0.00 \t\t 1.00 \t\t 0.00 \t\t 1.00\n",
      "Decision Tree                       0.64 \t\t 0.99 \t\t 0.31 \t\t 0.99\n",
      "Logistic Regression Under Sample    0.13 \t\t 0.85 \t\t 0.01 \t\t 0.92\n",
      "Decision Tree Under Sample          0.98 \t\t 0.93 \t\t 0.12 \t\t 0.97\n",
      "Logistic Regression Over Sample     0.09 \t\t 0.87 \t\t 0.01 \t\t 0.93\n",
      "Decision Tree Over Sample           0.25 \t\t 1.00 \t\t 0.28 \t\t 1.00\n",
      "Logistic Regression SMOTE           0.44 \t\t 0.48 \t\t 0.01 \t\t 0.65\n",
      "Decision Tree SMOTE                 0.56 \t\t 0.99 \t\t 0.38 \t\t 1.00\n",
      "Logistic Regression Near Miss       0.35 \t\t 0.63 \t\t 0.01 \t\t 0.77\n",
      "Decision Tree Near Miss             0.94 \t\t 0.70 \t\t 0.03 \t\t 0.82\n"
     ]
    }
   ],
   "source": [
    "getComparitiveChart(model_list,'Decision Tree Near Miss',decision_tree_model_nm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.5 Comparisions of different models\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "========================== Model Test Results ========================\n",
      "\n",
      "=== dc ===\n",
      "Model Accuracy:  98.7%\n",
      "\n",
      "Recall:\n",
      " 0.6384615384615384\n",
      "\n",
      "\n",
      "Classification Report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.99      0.99     27656\n",
      "           1       0.20      0.64      0.31       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.60      0.81      0.65     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n",
      "\n",
      "=== dc_us ===\n",
      "Model Accuracy:  93.30000000000001%\n",
      "\n",
      "Recall:\n",
      " 0.9769230769230769\n",
      "\n",
      "\n",
      "Classification Report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.93      0.97     27656\n",
      "           1       0.06      0.98      0.12       130\n",
      "\n",
      "    accuracy                           0.93     27786\n",
      "   macro avg       0.53      0.95      0.54     27786\n",
      "weighted avg       1.00      0.93      0.96     27786\n",
      "\n",
      "\n",
      "=== dc_os ===\n",
      "Model Accuracy:  99.4%\n",
      "\n",
      "Recall:\n",
      " 0.25384615384615383\n",
      "\n",
      "\n",
      "Classification Report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.31      0.25      0.28       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.65      0.63      0.64     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n",
      "\n",
      "=== dc_smote ===\n",
      "Model Accuracy:  99.2%\n",
      "\n",
      "Recall:\n",
      " 0.5615384615384615\n",
      "\n",
      "\n",
      "Classification Report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.99      1.00     27656\n",
      "           1       0.29      0.56      0.38       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.64      0.78      0.69     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n",
      "\n",
      "=== dc_nm ===\n",
      "Model Accuracy:  69.69999999999999%\n",
      "\n",
      "Recall:\n",
      " 0.9384615384615385\n",
      "\n",
      "\n",
      "Classification Report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.70      0.82     27656\n",
      "           1       0.01      0.94      0.03       130\n",
      "\n",
      "    accuracy                           0.70     27786\n",
      "   macro avg       0.51      0.82      0.42     27786\n",
      "weighted avg       0.99      0.70      0.82     27786\n",
      "\n",
      "\n",
      "============================= ROC Curve ===============================\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Test models\n",
    "classdict = {'normal':0, 'fraudulent':1}\n",
    "print()\n",
    "print('========================== Model Test Results ========================' \"\\n\")   \n",
    "modlist = [('dc', decision_tree_model),\n",
    "           ('dc_us', decision_tree_model_undersample),\n",
    "           ('dc_os', decision_tree_model_over_sample),\n",
    "           ('dc_smote', decision_tree_model_smote),\n",
    "           ('dc_nm', decision_tree_model_nm)\n",
    "           \n",
    "          ] \n",
    "models = [j for j in modlist]\n",
    "for i, v in models:\n",
    "    accuracy = metrics.accuracy_score(Y_test, v.predict(df_Test))\n",
    "    confusion_matrix = metrics.confusion_matrix(Y_test, v.predict(df_Test))\n",
    "    classification = metrics.classification_report(Y_test, v.predict(df_Test))   \n",
    "    print('=== {} ==='.format(i))\n",
    "    print (\"Model Accuracy: \",  '{}%'.format(np.round(accuracy, 3) * 100))\n",
    "    print()\n",
    "    print(\"Recall:\" \"\\n\", confusion_matrix[1][1]/130)\n",
    "    print()\n",
    "    #pf.plot_confusion_matrix(confusion_matrix, classes = list(classdict.keys()),title='Confusion Matrix Plot', cmap=plt.cm.summer)\n",
    "    print() \n",
    "    print(\"Classification Report:\" \"\\n\", classification) \n",
    "    print() \n",
    "\n",
    "print('============================= ROC Curve ===============================' \"\\n\")      \n",
    "pf.plot_roc_auc(arg1=models, arg2=df_Test, arg3=Y_test)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.5 Implementing the models\n",
    ">4.5.1\tLogistic Regression   \n",
    "4.5.2\tSVM    \n",
    "4.5.3\tXGBoost    \n",
    "4.5.4\tRandom Forest   \n",
    "4.5.5\tArtificial Neural Network   \n",
    "4.5.6   Penalized-SVM   \n",
    "\n",
    "\n",
    "Penalize Algorithms (Cost-Sensitive Training)\n",
    "The next tactic is to use penalized learning algorithms that increase the cost of classification mistakes on the minority class.\n",
    "\n",
    "A popular algorithm for this technique is Penalized-SVM.\n",
    "\n",
    "During training, we can use the argument class_weight=’balanced’ to penalize mistakes on the minority class by an amount proportional to how under-represented it is.\n",
    "\n",
    "We also want to include the argument probability=True if we want to enable probability estimates for SVM algorithms.\n",
    "\n",
    "Let’s train a model using Penalized-SVM on the original imbalanced dataset:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.6 Taking into consideration all samples that were considered fraud by any of the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.67      0.80     27656\n",
      "           1       0.01      1.00      0.03       130\n",
      "\n",
      "    accuracy                           0.68     27786\n",
      "   macro avg       0.51      0.84      0.42     27786\n",
      "weighted avg       1.00      0.68      0.80     27786\n",
      "\n",
      "\u001b[1m  Recall of combined results is : 1.0\n"
     ]
    }
   ],
   "source": [
    "# method to combine all the fraud samples\n",
    "modlist = [('dc      ', decision_tree_model),\n",
    "           ('dc_us   ', decision_tree_model_undersample),\n",
    "           ('dc_os   ', decision_tree_model_over_sample),\n",
    "           ('dc_smote', decision_tree_model_smote),\n",
    "           ('dc_nm   ', decision_tree_model_nm)]\n",
    "res_recall = {}\n",
    "res = pd.DataFrame()\n",
    "for i,v in modlist:\n",
    "    pred = v.predict(df_Test)\n",
    "    y_df = pd.DataFrame(pred)\n",
    "    res = pd.concat([res,y_df],axis = 1)\n",
    "    #recall\n",
    "    prfs = precision_recall_fscore_support(Y_test,pred , average=None)\n",
    "    res_recall[i] = round(prfs[1][1],2)\n",
    "\n",
    "# column total has value 1 if any of the models has predicted a fraud\n",
    "res.loc[:,'total'] = res.max(axis = 1)  \n",
    "res = res['total']\n",
    "report = classification_report(Y_test,res)\n",
    "print(classification_report(Y_test,res))\n",
    "\n",
    "prfs = precision_recall_fscore_support(Y_test,res , average=None)\n",
    "res_recall['Combined'] = round(prfs[1][1],2)\n",
    "print('\\033[1m', \" Recall of combined results is :\" ,round(prfs[1][1],4) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model            Recall\n",
      "dc       \t 0.64\n",
      "dc_us    \t 0.98\n",
      "dc_os    \t 0.25\n",
      "dc_smote \t 0.56\n",
      "dc_nm    \t 0.94\n",
      "Combined \t 1.0\n"
     ]
    }
   ],
   "source": [
    "# report generation of recall of all models\n",
    "print(\"Model\" , \"           Recall\")\n",
    "for i,v in res_recall.items():\n",
    "    print( i ,\"\\t\",v)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Implementing models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(129668, 3077) (27786, 3077) (129668,) (27786,)\n"
     ]
    }
   ],
   "source": [
    "# files ready for testing on models\n",
    "print(df_Train.shape, df_Test.shape, Y_train.shape, Y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest Classifier\n",
    "\n",
    "#### sklearn default parameters:    \n",
    "\n",
    "class sklearn.ensemble.RandomForestClassifier(n_estimators=100, *, criterion='gini', max_depth=None, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0.0, max_features='auto', max_leaf_nodes=None, min_impurity_decrease=0.0, min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=None, random_state=None, verbose=0, warm_start=False, class_weight=None, ccp_alpha=0.0, max_samples=None)    \n",
    "\n",
    "\n",
    "The Random forest or Random Decision Forest is a supervised Machine learning algorithm used for classification, regression, and other tasks using decision trees.\n",
    "The Random forest classifier creates a set of decision trees from a randomly selected subset of the training set. It is basically a set of decision trees (DT) from a randomly selected subset of the training set and then It collects the votes from different decision trees to decide the final prediction.\n",
    "\n",
    "Folowing are the hyper parameters to be tuned for Random Forest :   \n",
    ">- max_depth - The max_depth of a tree in Random Forest is defined as the longest path between the root node and the leaf node.   \n",
    "- min_sample_split – a parameter that tells the decision tree in a random forest the minimum required number of observations in any given node in order to split it.    \n",
    "The default value of the minimum_sample_split is assigned to 2.  \n",
    "By increasing the value of the min_sample_split, we can reduce the number of splits that happen in the decision tree and therefore prevent the model from overfitting.    \n",
    "- max_leaf_nodes - This hyperparameter sets a condition on the splitting of the nodes in the tree and hence restricts the growth of the tree. If after splitting we have more terminal nodes than the specified number of terminal nodes, it will stop the splitting and the tree will not grow further.   \n",
    "- min_samples_leaf - This hyperparameter sets a condition on the splitting of the nodes in the tree and hence restricts the growth of the tree. If after splitting we have more terminal nodes than the specified number of terminal nodes, it will stop the splitting and the tree will not grow further.   \n",
    "- n_estimators - Random Forest algorithm is a grouping of trees. n_estimators is the number of trees.More trees should be able to produce a more generalized result, but by choosing more number of trees, the time complexity of the Random Forest model also increases.      \n",
    "- max_sample (bootstrap sample) - The max_samples hyperparameter determines what fraction of the original dataset is given to any individual tree.    \n",
    "- max_features -  It is the number of maximum features provided to each tree in a random forest.  \n",
    "default=”auto”"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'print_eval' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-31-325c2f85b83b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     11\u001b[0m \u001b[0my_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel_random_forest\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf_Test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     12\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 13\u001b[0;31m \u001b[0mprint_eval\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_pred\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mmodel_random_forest\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'print_eval' is not defined"
     ]
    }
   ],
   "source": [
    "# raw data max_leaf_nodes = None(unlimited)\n",
    "# creating a RF classifier \n",
    "from sklearn.metrics import confusion_matrix\n",
    "model_random_forest = RandomForestClassifier(n_estimators = 100,random_state=137)   \n",
    "  \n",
    "# Training the model on the training dataset \n",
    "# fit function is used to train the model using the training sets as parameters \n",
    "model_random_forest.fit(df_Train, Y_train) \n",
    "\n",
    "# performing predictions on the test dataset \n",
    "y_pred = model_random_forest.predict(df_Test) \n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9999922879970385\n",
      "Testing Accuracy:  0.9963290865903692\n",
      "[[27655     1]\n",
      " [  101    29]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.97      0.22      0.36       130\n",
      "\n",
      "    accuracy                           1.00     27786\n",
      "   macro avg       0.98      0.61      0.68     27786\n",
      "weighted avg       1.00      1.00      1.00     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00    128963\n",
      "           1       1.00      1.00      1.00       705\n",
      "\n",
      "    accuracy                           1.00    129668\n",
      "   macro avg       1.00      1.00      1.00    129668\n",
      "weighted avg       1.00      1.00      1.00    129668\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# on train data\n",
    "y_pred11 = model_random_forest.predict(df_Train)\n",
    "print(classification_report(Y_train,y_pred11))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9999922879970385\n",
      "Testing Accuracy:  0.9963290865903692\n",
      "[[27656     0]\n",
      " [  130     0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.00      0.00      0.00       130\n",
      "\n",
      "    accuracy                           1.00     27786\n",
      "   macro avg       0.50      0.50      0.50     27786\n",
      "weighted avg       0.99      1.00      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# raw data max_leaf_nodes = None(unlimited)\n",
    "# creating a RF classifier \n",
    "from sklearn.metrics import confusion_matrix\n",
    "model_random_forest2 = RandomForestClassifier(n_estimators = 130,random_state=137, max_leaf_nodes=10)   \n",
    "  \n",
    "# Training the model on the training dataset \n",
    "# fit function is used to train the model using the training sets as parameters \n",
    "model_random_forest2.fit(df_Train, Y_train) \n",
    "\n",
    "# performing predictions on the test dataset \n",
    "y_pred2 = model_random_forest2.predict(df_Test) \n",
    "print_eval(y_pred2,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred22 = model_random_forest2.predict(df_Train) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[128963      0]\n",
      " [   665     40]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.99      1.00      1.00    128963\n",
      "           1       1.00      0.06      0.11       705\n",
      "\n",
      "    accuracy                           0.99    129668\n",
      "   macro avg       1.00      0.53      0.55    129668\n",
      "weighted avg       0.99      0.99      0.99    129668\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(confusion_matrix(Y_train,y_pred22))\n",
    "print(classification_report(Y_train,y_pred22))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.611520382278988"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9999922879970385\n",
      "Testing Accuracy:  0.9963290865903692\n",
      "[[27655     1]\n",
      " [  101    29]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.97      0.22      0.36       130\n",
      "\n",
      "    accuracy                           1.00     27786\n",
      "   macro avg       0.98      0.61      0.68     27786\n",
      "weighted avg       1.00      1.00      1.00     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# raw data \n",
    "# creating a RF classifier \n",
    "from sklearn.metrics import confusion_matrix\n",
    "model_random_forest = RandomForestClassifier(random_state=137)   \n",
    "  \n",
    "# Training the model on the training dataset \n",
    "# fit function is used to train the model using the training sets as parameters \n",
    "model_random_forest.fit(df_Train, Y_train) \n",
    "\n",
    "# performing predictions on the test dataset \n",
    "y_pred = model_random_forest.predict(df_Test) \n",
    "\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9999922879970385\n",
      "Testing Accuracy:  0.9963290865903692\n",
      "[[27641    15]\n",
      " [  127     3]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.17      0.02      0.04       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.58      0.51      0.52     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(random_state=137) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5112672726463585"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "WE see that performance deteriorates on limiting the max_leaf_nodes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Working with SMOTE data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26861   795]\n",
      " [   44    86]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.98     27656\n",
      "           1       0.10      0.66      0.17       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.55      0.82      0.58     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 50 ,random_state=137, max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8163962194877727"
      ]
     },
     "execution_count": 163,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26955   701]\n",
      " [   42    88]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.99     27656\n",
      "           1       0.11      0.68      0.19       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.55      0.83      0.59     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 75 ,random_state=137, max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8257879775705925"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[27641    15]\n",
      " [  126     4]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.21      0.03      0.05       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.60      0.52      0.53     27786\n",
      "weighted avg       0.99      0.99      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# smote   max_leaf_nodes = None(unlimited)\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 100,random_state=137) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5151134264925123"
      ]
     },
     "execution_count": 167,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26953   703]\n",
      " [   41    89]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.99     27656\n",
      "           1       0.11      0.68      0.19       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.56      0.83      0.59     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote  max_leaf_nodes = 16\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 100,random_state=137, max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8295979728977994"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[27182   474]\n",
      " [   75    55]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.98      0.99     27656\n",
      "           1       0.10      0.42      0.17       130\n",
      "\n",
      "    accuracy                           0.98     27786\n",
      "   macro avg       0.55      0.70      0.58     27786\n",
      "weighted avg       0.99      0.98      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 100,random_state=137, max_leaf_nodes=32) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7029688925480074"
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Increasing the max_leaf_nodes does not give any improvement, so we proceed with value 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26948   708]\n",
      " [   38    92]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.99     27656\n",
      "           1       0.12      0.71      0.20       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.56      0.84      0.59     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 125,random_state=137, max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8410460381388932"
      ]
     },
     "execution_count": 173,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26956   700]\n",
      " [   38    92]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.99     27656\n",
      "           1       0.12      0.71      0.20       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.56      0.84      0.59     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 130,random_state=137, max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8411906722146815"
      ]
     },
     "execution_count": 175,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26981   675]\n",
      " [   41    89]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.98      0.99     27656\n",
      "           1       0.12      0.68      0.20       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.56      0.83      0.59     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 140,random_state=137, max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8301041921630583"
      ]
     },
     "execution_count": 177,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26953   703]\n",
      " [   39    91]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.99     27656\n",
      "           1       0.11      0.70      0.20       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.56      0.84      0.59     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 131,random_state=137, max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8372902805901071"
      ]
     },
     "execution_count": 179,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Continuing with n_estimators = 100,random_state=130, max_leaf_nodes=16**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26767   889]\n",
      " [   70    60]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.98     27656\n",
      "           1       0.06      0.46      0.11       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.53      0.71      0.55     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote  max_leaf_nodes = 16\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 130,\n",
    "                                                   random_state=137, \n",
    "                                                   max_leaf_nodes=16,\n",
    "                                                  max_features = 6) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.714696769097261"
      ]
     },
     "execution_count": 181,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26500  1156]\n",
      " [   75    55]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.96      0.98     27656\n",
      "           1       0.05      0.42      0.08       130\n",
      "\n",
      "    accuracy                           0.96     27786\n",
      "   macro avg       0.52      0.69      0.53     27786\n",
      "weighted avg       0.99      0.96      0.97     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote  max_features = 20\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 130,\n",
    "                                                   random_state=137, \n",
    "                                                   max_leaf_nodes=16,\n",
    "                                                  max_features = 20) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6906388375870587"
      ]
     },
     "execution_count": 183,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[23481  4175]\n",
      " [  104    26]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.85      0.92     27656\n",
      "           1       0.01      0.20      0.01       130\n",
      "\n",
      "    accuracy                           0.85     27786\n",
      "   macro avg       0.50      0.52      0.46     27786\n",
      "weighted avg       0.99      0.85      0.91     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote  max_features = 20\n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 130,\n",
    "                                                   random_state=137, \n",
    "                                                   max_leaf_nodes=16,\n",
    "                                                  max_features = 3) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.524519091698004"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26956   700]\n",
      " [   38    92]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.99     27656\n",
      "           1       0.12      0.71      0.20       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.56      0.84      0.59     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote best model \n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 130,\n",
    "                                                   random_state=137, \n",
    "                                                   max_leaf_nodes=16) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8411906722146815"
      ]
     },
     "execution_count": 187,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26824   832]\n",
      " [   33    97]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.98     27656\n",
      "           1       0.10      0.75      0.18       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.55      0.86      0.58     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote best model \n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 130,\n",
    "                                                   random_state=137, \n",
    "                                                   max_leaf_nodes=10) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8580349791949444"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9945630379122066\n",
      "Testing Accuracy:  0.9953213848700785\n",
      "[[26796   860]\n",
      " [   37    93]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.98     27656\n",
      "           1       0.10      0.72      0.17       130\n",
      "\n",
      "    accuracy                           0.97     27786\n",
      "   macro avg       0.55      0.84      0.58     27786\n",
      "weighted avg       0.99      0.97      0.98     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote best model \n",
    "\n",
    "# training and predictions : random forest \n",
    "random_forest_model_smote = RandomForestClassifier(n_estimators = 130,\n",
    "                                                   random_state=137, \n",
    "                                                   max_leaf_nodes=9) \n",
    "random_forest_model_smote.fit(X_sm,y_sm)\n",
    "y_pred = random_forest_model_smote.predict(df_Test)\n",
    "print_eval(y_pred,model_random_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8421441445450703"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(Y_test, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### best hyper parameters for Random Forest\n",
    "(n_estimators = 130,\n",
    " random_state=137, \n",
    " max_leaf_nodes=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "   \n",
    "### 5.  Considering the Fraud cases as anamoly and use anamoly detection methods such as:   \n",
    "> 5.1 Simple Statistical Methods : Metrics such as distribution, including mean, median, mode, and quantiles could be used to identify outliers since the definition of an anomalous data point is one that deviates by a certain standard deviation from the mean. \n",
    "       5.2 Density-Based Anomaly Detection : These include the k-nearest neighbors algorithm, Relative density of data based method known as local outlier factor (LOF) algorithm    \n",
    "       5.3 Clustering-Based Anomaly Detection : K-means algorithm  \n",
    "       5.4 Support Vector Machine-Based Anomaly Detection      \n",
    "       5.5 Isolation Forest : https://www.youtube.com/watch?v=TP3wdwD8JVY    \n",
    "       5.6 Using Auto Encoders : https://www.youtube.com/watch?v=S31E-ftRfQI\n",
    "   \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "prec = [1,2,3,4]\n",
    "reca = [2,3,1,4]\n",
    "f1sc = [.2,.7,.8,.9]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " We will deal with anomaly detection in a separate notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtAAAAHwCAYAAACPE1g3AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/d3fzzAAAACXBIWXMAAAsTAAALEwEAmpwYAAAymklEQVR4nO3deZgdZZn38e9NAgRICAHiAgHDJgo4Romgo0xwAYFX4R1ld4QoiviKG4viIJCgMDAuMAKCOOwIBBgFHEEYZJVhCxqRfTNAACEJJBBCQhLu94+qDiedTvo86T69JN/PdfXVp+o8Veeu53Q6v37OU1WRmUiSJElqzkq9XYAkSZLUnxigJUmSpAIGaEmSJKmAAVqSJEkqYICWJEmSChigJUmSpAIGaEnLJCL+NSL+s5v3+cOImBYRf+/O/S5jLTdFxJeWcdtzI+KH3V1TV0RERsSmvV3HkkTE5hExKSJeiYhv9HY9PSEito+IKb1dh6RyBmhpORIRkyPi9YhYt936P9cBamQT+2jqP/XMPD4zlylgLuF1NwQOBbbIzLd10z77dGjsK+o/FubU4fXliLgnIo6IiFUL9tHVvv4OcGNmDsnMn3VhP231jIuIefUxvRIRj0TEqRHx9oJ9LPMfUQ372CYiro6IGRHxYkTcFRFf6Mo+u1jPxyPioYiYHRE3RsQ7eqsWqT8zQEvLn78B+7QtRMR7gNW78wUiYmB37q+2ITA9M1/oI/WsaA7OzCHA26n+kNkbuDoioode/x3A/cuy4VLe/wn1Ma0N/DPwNuCekhDdFRHxIeAG4GZgU2Ad4KvAzj3x+h3Usy7wa+Aoqj6ZCEzojVqk/s4ALS1/LgD2a1jeHzi/sUFErBoRP46IpyLi+Yg4IyJWi4g1gGuA9SJiVv21Xj2ad3lEXBgRLwNj63UXNuzzIxHxv/VI29MRMbZev0tEPFCPAj4TEYe1LzgiPgH8T8Prnluv3zUi7q/3eVNEvLthm8kR8d2IuBd4tbMQXdd7WX0Mr0TEXyPinRHxvYh4oa55x3abbVKPGL4cEVdGxNoN+7ssIv4eETMj4paI2HIJrzssIv47IqZGxEv14xENz98UET+IiNvquq5r/ARhKf3a4XvYsN3hEfFcRDwbEV9cWt80ysxXM/MmYFfgQ8D/qfe3TUTcXtfxXD2au0r93C315n+p37+9Ojvudn10A/BR4NR6+3dGxNCIOL/e/smI+H5ErFS3H1v310kRMR0Y18kxzcvM+4G9gKlUfyAs9b2JiOOA7RpqOrVe/x/1+9A2Ur/dUl76R8B5mXliZk7Lyj2ZuecS+uGIiHi8/jl4ICL+ueG5TSPi5vrnbVpETKjXR90PL9Q1/TUitlpCPZ8B7s/MyzJzTt1v742Idy2t/yQtzgAtLX/uANaMiHdHxACqkcQL27U5AXgnMIpqZGx94OjMfJVqdOzZzBxcfz1bb7MbcDmwFvCrxp1F9THwNcApwPB6v5Pqp88CvlKPBG5FNSK3iMy8vt3rjo2IdwIXA9+q93k18Nu20FbbhyrgrZWZ85vom09T/YExDPgzcC3V78H1gWOBX7Rrvx/wRapR2flA49SCa4DNgLcAf6JdnzRYCTiHaoR1Q+A14NR2bfYFvlDvaxXgMOi0Xzt8D+vtdqr3sUNd4yeWUNsSZeZTVCOUbQFxAfBtYF2qYP1x4P/Vbf+pbvPe+v2b0ORxt73Wx4BbqUbBB2fmI/UxDwU2BsZQvReNUx+2BZ4A3goc1+QxLQCubDimJdaYmUe2q+ngepu7qfp8beAi4LKIGNT+tSJi9bqfLm+mttrjdW1DgfHAhfHmaPkPgOuofnZHUPUPwI7AP1H9LAwF9gSmL2H/WwJ/aVuo/70/Xq+XVMAALS2f2kahdwAeBJ5peyIiAjgQ+HZmvpiZrwDHUwXtpbk9M6/IzDcy87V2z+0LXJ+ZF9ejfdMzc1L93Dxgi4hYMzNfysw/NXkMewG/y8z/ycx5wI+B1YB/bGjzs8x8uoN6luTWzLy2DtuXUYXSE+r9XwKMjIi1GtpfkJn31UHjKGDP+o8SMvPszHwlM+fy5kje0PYvWPfFf2Xm7Lqvj6MKhI3OycxH6uO4lCqgwRL6tYn3cM96n221j2uyf9p7liooUo+c3pGZ8zNzMtUfG+2Po/S4O9Twh9/36j6eDPwE+HxjbZl5Sl1Ps+9/+2MqrjEzL6y3m5+ZPwFWBTbvoOkwqv9jn2u2sHpk+Nn639gE4FFgm/rpeVRBf73MnJOZf2xYPwR4FxCZ+WBmLuk1BwMz262bWW8vqYABWlo+XUAVvsbSbvoGVWhcnWou6IyImAH8vl6/NE8v5bkNqEayOvJZYBfgyfoj6A918jpt1gOebFvIzDfqGtZvsqaOPN/w+DVgWj0q2bYMVcjoaP9PAisD60bEgIg4of64/WVgct1mkZM3oRqJjIhf1NMQXgZuAdZqC+K1xquOzG6oYUn92tl7uF4HtS+L9YEX6+N4Zz3F4e/1cRxPB8fbpsnjXpJ1qfq6se4n6dp736bxmIprjIjDIuLBeirFDKpR34764SXgDapPL5oSEftFdSWStvd0q4Z9fwcI4K6opjV9ESAzb6AaNT8NeCEizoyINSNiw3hzGtaseh+zgDXbveyawCvN1iipYoCWlkOZ+STVyYS7UJ001GgaVVjcMjPXqr+GZmZbaMsl7XYpL/k0sMkSark7M3ejmp5wBdUIazOepRpxAxaOnG9Aw2h6JzV1hw0aHm9INdo3jeqPk92opkYMBUa2ldnBPg6lGqHcNjPXpPq4fUlt21tSv3b2Hj7XQe1FImIDYGuqaQwApwMPAZvVx/GvnRxDV457Gm+OuLbZkC6+9/Uc6k/z5jF1VmO22347qiC7JzAsM9eiGsFd7JgyczZwO9UfkM3U9g7gl8DBwDr1vu9r23dm/j0zv5yZ6wFfAX4e9VVPMvNnmbk1sAXVVI7DM/OphmlYbT8X9wPvbXjNNah+vpbp5E1pRWaAlpZfBwAfqz/CX6geyf0lcFJEvAUgItaPiE/WTZ4H1uloOsJS/Ar4RETsGREDI2KdiBgVEatExOciYmg9TeJlqlG5ZlwK/J+oLru1MlXYmQv8b0FdXfUvEbFFPZ/1WODyesR6SF3LdKqR4OOXso8hVGF3RlQnIR5T8Pod9msT7+GlVCd6ttXe9GvWo7JjqOYK30U197ztOF4GZtUnnX213abPU81X7vJx1318KXBcRAypw+UhLD6Xvyl1372bak7924CfNlljR8c0n+pExIERcTSLj+g2+g7V+3B4RKxT1/LeiLikg7ZrUAX2qXW7L1CNQLcdwx7x5kmYL9Vt34iID0TEtvW/kVeBOSz539hvgK0i4rP1vO2jgXsz86GlHIOkDhigpeVUZj6emROX8PR3gceAO+qPrq+nnsdZ/2d6MfBE/VHyek281lNUo92HUn08Pok3R7o+D0yuX+cg4HNN1v8w8C9UJ0tNoxo5/HRmvt7M9t3kAuBcqikWg4C2G3ycTzWl4BngAaoTN5fkZKq529Pqdr9v9sU76delvYfX1K97Q91msRM3O3BqRLxCFRpPBv4L2KkO61CdlLgv1cf9v2Txy5+NA86rf2b27Mpx175OFQifAP5IdcLe2YX72KuevjATuIrqD56t880TYzur8T+A3aO6QsfPqE46/T3wCNX7P4elTCXJzP8FPlZ/PRERLwJn8uYfJY1tH6Ca53071XvwHuC2hiYfAO6sj+cq4JuZ+QRVgP8lVah+sj7GHy2hnqlUI+LH1e23pfNzHyR1IDJb/QmoJEmStPxwBFqSJEkqYICWJEmSChigJUmSpAIGaEmSJKmAAVqSJEkqMLC3Cyi17rrr5siRI3u7DEmSJC3n7rnnnmmZudidevtdgB45ciQTJy7p0raSJElS94iIJzta7xQOSZIkqYABWpIkSSpggJYkSZIK9Ls50JIkSVrUvHnzmDJlCnPmzOntUvqlQYMGMWLECFZeeeWm2hugJUmS+rkpU6YwZMgQRo4cSUT0djn9SmYyffp0pkyZwkYbbdTUNk7hkCRJ6ufmzJnDOuusY3heBhHBOuusUzR6b4CWJElaDhiel11p3xmgJUmS1GUDBgxg1KhRbLXVVuyxxx7Mnj27y/s8+uijuf7665f4/BlnnMH555/f5dcpFZnZ4y/aFaNHj05vpCJJkvSmBx98kHe/+90Ll0/46cnMfWVmt+1/1SFDOeKQby21zeDBg5k1axYAn/vc59h666055JBDFj4/f/58Bg7su6ffte9DgIi4JzNHt2/bd49CkiRJy2TuKzMZN+aYbtvfuJvHF7XfbrvtuPfee7nppps46qijGDZsGA899BAPPvggRxxxBDfddBNz587la1/7Gl/5ylcAOPHEE7nwwgtZaaWV2HnnnTnhhBMYO3Ysn/rUp9h999054ogjuOqqqxg4cCA77rgjP/7xjxk3bhyDBw/msMMOY9KkSRx00EHMnj2bTTbZhLPPPpthw4ax/fbbs+2223LjjTcyY8YMzjrrLLbbbrsu9YcBWpIkSd1m/vz5XHPNNey0004A/OlPf+K+++5jo4024swzz2To0KHcfffdzJ07lw9/+MPsuOOOPPTQQ1x55ZXceeedrL766rz44ouL7HP69On85je/4aGHHiIimDFjxmKvu99++3HKKacwZswYjj76aMaPH8/JJ5+8sKa77rqLq6++mvHjxy91WkgznAMtSZKkLnvttdcYNWoUo0ePZsMNN+SAAw4AYJtttll4ebjrrruO888/n1GjRrHtttsyffp0Hn30Ua6//nq+8IUvsPrqqwOw9tprL7LvoUOHMmjQIA444AB+/etfL2zXZubMmcyYMYMxY8YAsP/++3PLLbcsfP4zn/kMAFtvvTWTJ0/u8rE6Ai1JkqQuW2211Zg0adJi69dYY42FjzOTU045hU9+8pOLtLn22muXuu+BAwdy11138Yc//IHLL7+cU089lRtuuKHp2lZddVWgOtFx/vz5TW+3JI5AS5IkqUd88pOf5PTTT2fevHkAPPLII7z66qvssMMOnHPOOQuv3NF+CsesWbOYOXMmu+yyCyeddBJ/+ctfFnl+6NChDBs2jFtvvRWACy64YOFodCs4Ai1JkqQe8aUvfYnJkyfz/ve/n8xk+PDhXHHFFey0005MmjSJ0aNHs8oqq7DLLrtw/PHHL9zulVdeYbfddmPOnDlkJj/96U8X2/d555238CTCjTfemHPOOadlx9Gyy9hFxNnAp4AXMnOrDp4P4D+AXYDZwNjM/FNn+/UydpIkSYvqC5ex6+/6ymXszgVOBZZ0deudgc3qr22B0+vvkiRJ6oLlPez2tpbNgc7MW4AXl9JkN+D8rNwBrBURb29VPZIkSVJ36M2TCNcHnm5YnlKvkyRJkvqsfnESYUQcCBwIsOGGG/ZaHd09n6jNwAELmL9gQLfvF2AB8xnQgrd51dVX5YjDj+j2/YL93Mh+XlSr+hla19et6mdoXV/bz4uyn9/UH/sZVozf0TvuuCPPPvtsp+2CJImulLZE1Z67f9+xUvD2t/WtSQq9GaCfATZoWB5Rr1tMZp4JnAnVSYStL61j3X1bzDbjbh7PMXuN6/b9AoyfMI5xU7p/3+NGdP8+29jPb7KfF9WqfobW9XWr+hla19f286Ls5zf1x36GFeN39LaDHuTZIet12m69V55lvWGdB+1l8exL6/Hs692/7/VW6fy4elpvTuG4CtgvKh8EZmbmc71YjyRJktSplo1AR8TFwPbAuhExBTgGWBkgM88Arqa6hN1jVJex+0KrapEkSVJrbbDBBmyy+SYsWLCA9Uasx7EnHcuQNYd02/5HjhzJxIkTWXfddRk8eDCzZs3qtn2XalmAzsx9Onk+ga+16vUlSZJWVDf+8gRmvzK32/Y3dMhKfOurOy+1zaBBg7jo6osAGHfoOC47/zK+ePAXu62GvqRfnEQoSZKk5s1+ZW63zikfP6FsX+95/3t49KFHAZjy5BROPPpEZrw4g0GDBnHkCUcycpORTJ86nRO+fwLPPFWdAvfdH36X9279Xg478DCef+555s6dy95j9+bgsQd323F0FwO0JEmSus2CBQu4+7a72XWvXQE47l+P43s//B4bbrQh9/35Pk486kROv+h0fjz+x7xv2/fxo1/8iAULFvDaq68BcNS/H8XQtYYyZ84c9t9tf/bddV/WW69vnUhogJYkSVKXzZkzh3132Zepz09lo002YtuPbMvsV2fz13v+yhFfe/Nyf/NenwfAxNsnMv4n4wEYMGAAg9ccDMCEcydw07U3AfD8c8/zt7/9ja222qpnD6YTBmhJkiR1Wdsc6DmvzeHr+32dy86/jE/t/ikGrzl44dzoztxzxz3cddtdnP3rsxm02iC+svdXmDu3++Zyd5fevIydJEmSljODVhvEYeMO41f/+SsGrTaI9TZYj+t/dz0AmckjDzwCwAf+8QNc/qvLgWrax6yXZzHr5VkMWXMIg1YbxOTHJ3Pfn+/rteNYGgO0JEmSutXmW27Opu/alOuuuo4fnPwDrrz0SvbdeV/22nEvbr7+ZgAOPfpQ7rn9HvbeaW8+/+nP88RjT/ChMR9iwYIF7PGJPTj1xFPZ6n19a+pGG6dwSJIkLWdWH7Jq8ZUzlmbokM7HXB999NFF7kR40lknLXx8ynmnLNZ+neHr8JNf/mSx9T8792eLLLfdiXDy5MkL1/XmNaDBAC1JkrTc+eiXj+hwfStv5b0icQqHJEmSVMAALUmSJBUwQEuSJEkFDNCSJElSAQO0JEmSVMCrcEiSJKnLzjrrLM4+72w22mwjpj0/jYfuf4ivHvpVPn/g53u7tG5ngJYkSVrOXHP6CbzxWvfdAnvV1VfiiIN3Xmqb8847j1MvOJWVV16Z5555jpuvu7nbXr8Z8+fPZ+DAnom2BmhJkqTlzBuvzWXclHHdtr9xI5a+r4O+/m889dRTfPML32TXPXZl3wP25bYbb1vqNvfccQ8/Oba6kUpEcOaEM1lj8Bqcd8Z5XHPFNawUK/Gh7T/Evx31b0yaNImDDjqI2bNns8kmm3D22WczbNgwtt9+e0aNGsUf//hH9tlnH7bffnsOOeQQZs2axbrrrsu5557L29/+9u7qhoUM0JIkSeqSM075Hr+79i7OuOgM1lp7raa2ufCXF/LdY7/Le0e/l9mvzmaVVVfhtptu4+b/uZlzf3Mug1YbxMwZMwHYb7/9OOWUUxgzZgxHH30048eP5+STTwbg9ddfZ+LEicybN48xY8Zw5ZVXMnz4cCZMmMCRRx7J2Wef3e3Ha4CWJElSj3vv6Pdy0g9PYqfdduKjO32Ut779rdz1x7v49O6fZtBqgwAYutZQXn75ZWbMmMGYMWMA2H///dljjz0W7mevvfYC4OGHH+a+++5jhx12AGDBggUtGX0GA7QkSZJ6wKXnX8oVl1wBwH+c8x+M/epYPvLRj3DbTbfxpd2/xCnnn7JM+11jjTUAyEy23HJLbr/99u4qeYm8jJ0kSZJabs/99uSiqy/ioqsvYvhbhzPlySls+q5N2f+g/dniH7Zg8uOT2Xa7bfnt5b9lzmtzAJg5YyZrrrkmw4YN49ZbbwXgggsuWDga3WjzzTdn6tSpCwP0vHnzuP/++1tyLI5AS5IkqdtMmzqN/Xfdn1dnvUpEcMk5lzDhugkMHjJ4kXYXn30xE++YyEqxEhu/c2P+ccw/ssqqq/DIA4+w3677MXDlgXz4ox/muCOP47zzzlt4EuHGG2/MOeecs9jrrrLKKlx++eV84xvfYObMmcyfP59vfetbbLnllt1+jAZoSZKk5cxKq63a6ZUzSqy6eueTFu68806eff1ZAH53++86bX/4+MM7XD/2q2MZ+9Wxi6wbNWoUd9xxx2Jtb7rppsXa3XLLLZ2+dlcZoCVJkpYzO3/1iA7Xr/fKs6w37Nkermb54xxoSZIkqYABWpIkSSpggJYkSern3gDI7O0y+q0s7DsDtCRJUj/32IJBzJ853RC9DDKT6dOnM2jQoKa38SRCSZKkfm7cqyMY9/wUNp02damjo/PmzGDmtJktqWHGq/OYuaD79z1vwDxmzmxNzW0GDRrEiBEjmm5vgJYkSernXsqV+easjTptN+7m8Ryz17iW1DD+N+MYN6X79z1uxDiOOeaYbt9vVziFQ5IkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSrQ0gAdETtFxMMR8VhEHNHB8xtGxI0R8eeIuDcidmllPZIkSVJXtSxAR8QA4DRgZ2ALYJ+I2KJds+8Dl2bm+4C9gZ+3qh5JkiSpO7RyBHob4LHMfCIzXwcuAXZr1yaBNevHQ4FnW1iPJEmS1GWtDNDrA083LE+p1zUaB/xLREwBrga+3tGOIuLAiJgYEROnTp3ailolSZKkpvT2SYT7AOdm5ghgF+CCiFispsw8MzNHZ+bo4cOH93iRkiRJUptWBuhngA0alkfU6xodAFwKkJm3A4OAdVtYkyRJktQlrQzQdwObRcRGEbEK1UmCV7Vr8xTwcYCIeDdVgHaOhiRJkvqslgXozJwPHAxcCzxIdbWN+yPi2IjYtW52KPDliPgLcDEwNjOzVTVJkiRJXTWwlTvPzKupTg5sXHd0w+MHgA+3sgZJkiSpO/X2SYSSJElSv2KAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgq0NEBHxE4R8XBEPBYRRyyhzZ4R8UBE3B8RF7WyHkmSJKmrBjbTKCIC+H/ADsAP6u93ZeYNS9lmAHBa3XYKcHdEXJWZDzS02Qz4HvDhzHwpIt6yzEciSZIk9YBmR6CPB04BPg0MBd4NjO9km22AxzLzicx8HbgE2K1dmy8Dp2XmSwCZ+UKzhUuSJEm9odkAvR/wi4blPwLv6WSb9YGnG5an1OsavRN4Z0TcFhF3RMROHe0oIg6MiIkRMXHq1KlNlixJkiR1v2YD9GrAcw3L6wPzuuH1BwKbAdsD+wC/jIi12jfKzDMzc3Rmjh4+fHg3vKwkSZK0bJqaAw3cCBxSP/4x1ejzrzvZ5hlgg4blEfW6RlOAOzNzHvC3iHiEKlDf3WRdkiRJUo9qdgT668Cf68ejgFuBb3eyzd3AZhGxUUSsAuwNXNWuzRVUo89ExLpUUzqeaLImSZIkqcc1NQKdmc8CH42INerlV5vYZn5EHAxcCwwAzs7M+yPiWGBiZl5VP7djRDwALAAOz8zpy3gskiRJUss1exm784EnMnNcvTwe2Cgz91vadpl5NXB1u3VHNzxOqqkhhyBJkiT1A81O4fgs8GTD8pPAZ7q/HEmSJKlvazZAzwDGNCxvD8zs7mIkSZKkvq7Zq3D8FjgwIj5ZL78FOLM1JUmSJEl9V7MB+nBgFeBT9fK5wHdaUZAkSZLUlzV7FY5XgC+2uBZJkiSpz2v2KhwfBsYBI6kuSQfVRTQ2aU1ZkiRJUt/U7BSOi6nuJDgXmN+6ciRJkqS+rdmrcAB8PzNXy8whbV8tq0qSJEnqo5odgb4C2CUi7gRealuZmX9qRVGSJElSX9VsgD4YSOC6dusHdNBWkiRJWm41G6DPpwrQkiRJ0gqt2cvYjW1xHZIkSVK/0Oxl7NYAvgG8BxhUr87M/GyrCpMkSZL6omancPwnsBfVNI6o1zmlQ5IkSSucZi9j9wng1PrxXsDlwPdbUpEkSZLUhzUboAcD91KNPq8O3A0c1qqiJEmSpL6q2SkcU6hC9OPAWVRB+ulWFSVJkiT1Vc0G6IOAWcAk4ASq+c//2qKaJEmSpD6r2cvY/aFh8YMtqkWSJEnq85qaAx0RO0bEnyLipYh4uf6a2eriJEmSpL6m2Skc5wJvAZ4FFrSsGkmSJKmPazZAzwe+lZmndtpSkiRJWo41G6APBH4eEesCL9frMjNPak1ZkiRJUt/UbIA+GBgJHN2wLgEDtCRJklYozQbojwC/B34NzGtdOZIkSVLf1myAvpBqxPnczJzfwnokSZKkPq3ZAL0/1S28D4qI1+p1mZlDW1OWJEmS1Dc1G6CnA9NaWYgkSZLUH3QaoCNiJWBXYGpmPtf6kiRJkqS+q9M7EWbmG8DNwM6tL0eSJEnq25q6lTdwEfCpiBjcymIkSZKkvq7kJMLVgJkR8Wq9zpMIJUmStMJpNkBPo7qMnSRJkrRCaypAZ+bIFtchSZIk9QtNzYGOiJUjYlxE3BkRd0TE0RGxcquLkyRJkvqaZqdw/DvwTeCNevkDwFrAIS2oSZIkSeqzmr0Kx57AOVR3I1wDOBfYq0U1SZIkSX1WswF6NeDhzHw9M+cCj9TrJEmSpBVKs1M4bgGOi4hPU12N44PAf7esKkmSJKmPajZAHwysDXykXr4F+HpLKpIkSZL6sKUG6Ig4FpgAbA3sB0wFyMxXl7adJEmStLzqbA70kcD7qE4g/EBmvmp4liRJ0oqsswA9HTgNCODUiHii4evx1pcnSZIk9S2dBejjgbn14zWB4Q1fb2lhXZIkSVKftNQAnZknZ+ZbgJuBnTNzSONXz5QoSZIk9R2dXgc6IgYA7wY2bX05kiRJUt/WaYDOzAXAX4FNWl+OJEmS1Lc1ex3o1YHvRMQOwLP1uszM3VpTliRJktQ3NRugP1R/f3/9BdUdCSVJkqQVSrMBeqOWViFJkiT1E53OgQbIzCeBocA/AwOADYE3WliXJEmS1Cc1NQIdEXsDF1LdUOVe4HvALKpALUmSJK0wmhqBBsYDf2hY/h3wj91fjiRJktS3NRug1wNuaFieB6zW/eVIkiRJfVuzJxHeB+xXP/48sBPwl5ZUJEmSJPVhzY5Afxt4K9Uc6P2BlYHDWlWUJEmS1FctdQQ6IoYB/w5sC9wB3EY18nx7Zr7U+vIkSZKkvqWzKRw/B/aqH28FvAfYNDPntbQqSZIkqY/qbArHjsB/AVsARwIj6seSJEnSCqmzAD0MuDwzHwJ+STUHeljLq5IkSZL6qGauwnF4RHyO6sTBBI6PiGlAZuZuLa1OkiRJ6mOaCdDvr7/afLD+nt1fjiRJktS3dRagN+qRKiRJkqR+YqkBOjOf7KlCJEmSpP6g2RupSJIkScIALUmSJBUxQEuSJEkFDNCSJElSAQO0JEmSVMAALUmSJBUwQEuSJEkFDNCSJElSAQO0JEmSVKClAToidoqIhyPisYg4YintPhsRGRGjW1mPJEmS1FUtC9ARMQA4DdgZ2ALYJyK26KDdEOCbwJ2tqkWSJEnqLq0cgd4GeCwzn8jM14FLgN06aPcD4ERgTgtrkSRJkrpFKwP0+sDTDctT6nULRcT7gQ0y83ctrEOSJEnqNr12EmFErAT8FDi0ibYHRsTEiJg4derU1hcnSZIkLUErA/QzwAYNyyPqdW2GAFsBN0XEZOCDwFUdnUiYmWdm5ujMHD18+PAWlixJkiQtXSsD9N3AZhGxUUSsAuwNXNX2ZGbOzMx1M3NkZo4E7gB2zcyJLaxJkiRJ6pKWBejMnA8cDFwLPAhcmpn3R8SxEbFrq15XkiRJaqWBrdx5Zl4NXN1u3dFLaLt9K2uRJEmSuoN3IpQkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAi0N0BGxU0Q8HBGPRcQRHTx/SEQ8EBH3RsQfIuIdraxHkiRJ6qqWBeiIGACcBuwMbAHsExFbtGv2Z2B0Zv4DcDnw762qR5IkSeoOrRyB3gZ4LDOfyMzXgUuA3RobZOaNmTm7XrwDGNHCeiRJkqQua2WAXh94umF5Sr1uSQ4ArmlhPZIkSVKXDeztAgAi4l+A0cCYJTx/IHAgwIYbbtiDlUmSJEmLauUI9DPABg3LI+p1i4iITwBHArtm5tyOdpSZZ2bm6MwcPXz48JYUK0mSJDWjlQH6bmCziNgoIlYB9gauamwQEe8DfkEVnl9oYS2SJElSt2hZgM7M+cDBwLXAg8ClmXl/RBwbEbvWzX4EDAYui4hJEXHVEnYnSZIk9QktnQOdmVcDV7dbd3TD40+08vUlSZKk7uadCCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpgAFakiRJKmCAliRJkgoYoCVJkqQCBmhJkiSpQEsDdETsFBEPR8RjEXFEB8+vGhET6ufvjIiRraxHkiRJ6qqWBeiIGACcBuwMbAHsExFbtGt2APBSZm4KnASc2Kp6JEmSpO7QyhHobYDHMvOJzHwduATYrV2b3YDz6seXAx+PiGhhTZIkSVKXtDJArw883bA8pV7XYZvMnA/MBNZpYU2SJElSl0RmtmbHEbsDO2Xml+rlzwPbZubBDW3uq9tMqZcfr9tMa7evA4ED68XNgYdbUnT/sC4wrdNW6g72dc+wn3uG/dwz7OeeYT/3DPsZ3pGZw9uvHNjCF3wG2KBheUS9rqM2UyJiIDAUmN5+R5l5JnBmi+rsVyJiYmaO7u06VgT2dc+wn3uG/dwz7OeeYT/3DPt5yVo5heNuYLOI2CgiVgH2Bq5q1+YqYP/68e7ADdmqIXFJkiSpG7RsBDoz50fEwcC1wADg7My8PyKOBSZm5lXAWcAFEfEY8CJVyJYkSZL6rFZO4SAzrwaubrfu6IbHc4A9WlnDcsipLD3Hvu4Z9nPPsJ97hv3cM+znnmE/L0HLTiKUJEmSlkfeyluSJEkqYIDuwyJiVgfrxkXEMxExKSIeiIh9eqO25UVELKj78v6I+EtEHBoRKzU8v01E3FLfkv7PEfGfEbF6b9bcX0TEkXW/3lv38TER8W/t2oyKiAfrx5Mj4tZ2z0+qL3epAg0/1/dFxG8jYq16/ciIeK1+ru1rlV4ut1+KiLdGxEUR8URE3BMRt0fEP0fE9hExs+7beyPi+oh4S2/X219ExAYR8beIWLteHlYvj4yIzSLivyPi8brPb4yIf6rbjY2IqQ2/zy/3d3XnIuL/RkRGxLvq5ZH18tcb2pwaEWPrx+fWGWTVenndiJjcG7X3NgN0/3RSZo6iupPjLyJi5V6upz97LTNHZeaWwA5Ut54/Bqr/IIHLgO9m5uaZ+T7g98CQXqu2n4iIDwGfAt6fmf8AfAK4EdirXdO9gYsblodExAb1Pt7dE7Uup9p+rreiOkH7aw3PPV4/1/b1ei/V2G/Vd8y9ArglMzfOzK2pfpZH1E1urfv2H6iuSPW1jvek9jLzaeB04IR61QlU83D/DvwOODMzN6n7/OvAxg2bT2j4ff46i/++0eL2Af5Yf2/zAvDNpfxxvQD4YqsL6+sM0P1YZj4KzAaG9XYty4PMfIHqhj0H1/9Bfg04LzNvb2hzeWY+31s19iNvB6Zl5lyAzJyWmbcAL0XEtg3t9mTRAH0pb/6nt0+757Rsbmfxu8Cqaz4GvJ6ZZ7StyMwnM/OUxkb175EhwEs9XF9/dxLwwYj4FvAR4MfA54Db6yt4AZCZ92Xmue03ru8rsQb2+1JFxGCq/j2ARa+CNhX4A29eZri9k4Fv1/28wjJA92MR8X7g0Tr4qRtk5hNUl118C7AVcE/vVtRvXQdsEBGPRMTPI2JMvf5i6l/UEfFB4MX6D8E2/wV8pn78aeC3PVXw8igiBgAfZ9Fr8G/SMH3jtF4qrb/bEvjTUp7fLiImAU9Rffpydk8UtbzIzHnA4VRB+lv1cmd9DrBX3e/PAGvj74/O7Ab8PjMfAaZHxNYNz50IHFb/DmnvKapR68/3QI19lgG6f/p2RNwP3Akc19vFSO1l5ixga6oR/anAhHoO3QRg93qeefvpG1DdifSliNgbeJDqExaVW60OEn8H3gr8T8NzjVM4nFrQDSLitPocirvrVW1TODYAzgH+vRfL6692Bp6jGshYTET8pp7j/+uG1RPq6Y1vA/5KFcK1ZPsAl9SPL6FhGkc9mHQnsO8Stv03qv5dYXPkCnvg/dxJ9RyvzwJnRcSg3i5oeRERG1PN73oBuJ8qBGoZZOaCzLwpM48BDgY+W89v/Bswhurnd0IHm04ATsPpG13xWh0k3gG0TUdS97kfeH/bQv2HyMeB4R20vQr4px6qa7kQEaOozkn5INWA0dtZvM//GRhLNdK8iPqOxr/Ffl+i+iTNjwH/WZ8EeDjVlLpoaHY88N1264CFU0gn1duskAzQ/Vg9F2wiS56npAIRMRw4Azi1/gV8KrB/45zdiPhMfXKhliIiNo+IzRpWjQKerB9fTPXR7BOZOaWDzX9DNWJ3bUuLXAFk5mzgG8ChK/p8xW52AzAoIr7asG5JV3z4CPB460taPtTzxk+nmrrxFPAjqjnQFwEfjohdG5ov7Sob9vvS7Q5ckJnvyMyR9aclfwM2aGuQmQ8BD1BNp+vIccBhLa+0j/IXat+2ekQ0BoyfdtDmWOCiiPhlZr7RQ3UtT9o+6l4ZmA9cQN3Pmfl8PZXgx/VlqN4AbqG6EoeWbjBwSn35tPnAY1TTOaC6ssnPqM6gX0xmvkI1/47q/1J1RWb+OSLupfp49tbO2qtzmZkR8X+BkyLiO1TTlF6lGq2DN+dABzAT+FJv1NlPfRl4KjPbph39HPgCsA3VlX1+GhEnA88DrwA/bNh2r4j4CNXg4BSqEWp1bB/q37MN/gv4Xrt1xwF/7mgHmXl/RPyJhk8GViTeiVCSJEkq4BQOSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpL6gYgYGRFZfx3VsP6stvWF+7uvmW0iYly9/92XpW5JWh4ZoCWp/xkblcGswHcCk6TeYoCWpP7lCWBjYHuq8Lwy8AxUd3GLiO9HxJMR8UpE3BgRW9bPrRURv4uIVyPignq7hSLiexHxt3q7a+vb2kuSOmCAlqT+5UHgTuCL9dcVwIz6uS8APwDuBY4EPgBcGRErA8cAuwC/Ap4C3tm2w4jYHzi+3u8JwD9Q3TFSktQBb+UtSf3P2VS3Q18V2An4Sb1+l/r7IZn5aERsC+xLFZa3p7od/cGZ+XpE7AeMqNt/qv6+V/0F8LaIWLulRyFJ/ZQBWpL6n0uAk4ApwP908Hy2+96szwEv1I9XAmYvU3WStJxzCock9TOZ+TLV9I2vZOYbDU/9rv7+04j4OrAb8DjwCHAj1e/8UyPiON4cfQb47/r7/sAGwBjgqMyc07qjkKT+yxFoSeqHMnNCB6vPBdYHvgx8DLibasrGvIj4AfAuqika1wCPAZvW+zovIt4GfAU4nWpku6P9S5KAyCz9hE+SJElacTmFQ5IkSSpggJYkSZIKGKAlSZKkAgZoSZIkqYABWpIkSSpggJYkSZIKGKAlSZKkAgZoSZIkqcD/B41UMYcK3pORAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# for raw class 0\n",
    "\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt \n",
    "   \n",
    "# set width of bar \n",
    "barWidth = 0.25\n",
    "fig = plt.subplots(figsize =(12, 8)) \n",
    "   \n",
    "# set height of bar \n",
    "prec =  [1 ,1,1,1,1,1,1]\n",
    "recall= [1 ,1,1,1,1,1,1]\n",
    "f1score = [1 ,1,1,1,1,1,1]\n",
    "   \n",
    "# Set position of bar on X axis \n",
    "br1 = np.arange((7)) \n",
    "br2 = [x + barWidth for x in br1] \n",
    "br3 = [x + barWidth for x in br2] \n",
    "   \n",
    "# Make the plot \n",
    "plt.bar(br1, prec, color ='deepskyblue', width = barWidth, \n",
    "        edgecolor ='grey', label ='Precision') \n",
    "plt.bar(br2, recall, color ='gold', width = barWidth, \n",
    "        edgecolor ='grey', label ='Recall') \n",
    "plt.bar(br3, f1score, color ='darkgreen', width = barWidth, \n",
    "        edgecolor ='grey', label ='f1-score') \n",
    "   \n",
    "# Adding Xticks  \n",
    "plt.xlabel('Model', fontweight ='bold') \n",
    "plt.ylabel('Performance', fontweight ='bold') \n",
    "plt.xticks([r + barWidth for r in range(len(prec))], \n",
    "           ['LR', 'DC', 'SVM', 'RF', 'GB','XGB','ANN']) \n",
    "plt.legend()  \n",
    "plt.title(\"Metrics for Imbalanced Data for Data Class-0\")\n",
    "plt.show() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# for smote class 0\n",
    "\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt \n",
    "   \n",
    "# set width of bar \n",
    "barWidth = 0.25\n",
    "fig = plt.subplots(figsize =(12, 8)) \n",
    "   \n",
    "# set height of bar \n",
    "prec =  [1 ,1,1,1,1,1,1]\n",
    "recall= [.95,.98,1,.99,.99,.99,1]\n",
    "f1score = [.97,.98,1,.99,.99,1,1]\n",
    "   \n",
    "# Set position of bar on X axis \n",
    "br1 = np.arange((7)) \n",
    "br2 = [x + barWidth for x in br1] \n",
    "br3 = [x + barWidth for x in br2] \n",
    "   \n",
    "# Make the plot \n",
    "plt.bar(br1, prec, color ='deepskyblue', width = barWidth, \n",
    "        edgecolor ='grey', label ='Precision') \n",
    "plt.bar(br2, recall, color ='gold', width = barWidth, \n",
    "        edgecolor ='grey', label ='Recall') \n",
    "plt.bar(br3, f1score, color ='darkgreen', width = barWidth, \n",
    "        edgecolor ='grey', label ='f1-score') \n",
    "   \n",
    "# Adding Xticks  \n",
    "plt.xlabel('Model', fontweight ='bold') \n",
    "plt.ylabel('Performance', fontweight ='bold') \n",
    "plt.xticks([r + barWidth for r in range(len(prec))], \n",
    "           ['LR', 'DC', 'SVM', 'RF', 'GB','XGB','ANN']) \n",
    "plt.legend()  \n",
    "plt.title(\"Metrics SMOTE-generated Data for Class-0\")\n",
    "plt.show() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# for raw class 1\n",
    "\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt \n",
    "   \n",
    "# set width of bar \n",
    "barWidth = 0.25\n",
    "fig = plt.subplots(figsize =(12, 8)) \n",
    "   \n",
    "# set height of bar \n",
    "prec = [0,.78,.46,.98,.73,.85,.38]   \n",
    "recall= [0,.63,.25,.47,.57,.65,.21]\n",
    "f1score = [0,.70,.32,.64,.64,.74,.27] \n",
    "   \n",
    "# Set position of bar on X axis \n",
    "br1 = np.arange((7)) \n",
    "br2 = [x + barWidth for x in br1] \n",
    "br3 = [x + barWidth for x in br2] \n",
    "   \n",
    "# Make the plot \n",
    "plt.bar(br1, prec, color ='deepskyblue', width = barWidth, \n",
    "        edgecolor ='grey', label ='Precision') \n",
    "plt.bar(br2, recall, color ='gold', width = barWidth, \n",
    "        edgecolor ='grey', label ='Recall') \n",
    "plt.bar(br3, f1score, color ='darkgreen', width = barWidth, \n",
    "        edgecolor ='grey', label ='f1-score') \n",
    "   \n",
    "# Adding Xticks  \n",
    "plt.xlabel('Model', fontweight ='bold') \n",
    "plt.ylabel('Performance', fontweight ='bold') \n",
    "plt.xticks([r + barWidth for r in range(len(prec))], \n",
    "           ['LR', 'DC', 'SVM', 'RF', 'GB','XGB','ANN']) \n",
    "plt.legend()  \n",
    "plt.title(\"Metrics for Imbalanced Data for Class-1\")\n",
    "plt.show() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# for raw class 1\n",
    "\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt \n",
    "   \n",
    "# set width of bar \n",
    "barWidth = 0.25\n",
    "fig = plt.subplots(figsize =(12, 8)) \n",
    "  \n",
    "# set height of bar \n",
    "prec = [.07,.15,.16,.23,.25,.36,.11]  \n",
    "recall= [.77,.76,.08,.62,.71,.78,.05]\n",
    "f1score = [.13,.25,.11,.33,.37,.49,.07]\n",
    "   \n",
    "# Set position of bar on X axis \n",
    "br1 = np.arange((7)) \n",
    "br2 = [x + barWidth for x in br1] \n",
    "br3 = [x + barWidth for x in br2] \n",
    "   \n",
    "# Make the plot \n",
    "plt.bar(br1, prec, color ='deepskyblue', width = barWidth, \n",
    "        edgecolor ='grey', label ='Precision') \n",
    "plt.bar(br2, recall, color ='gold', width = barWidth, \n",
    "        edgecolor ='grey', label ='Recall') \n",
    "plt.bar(br3, f1score, color ='darkgreen', width = barWidth, \n",
    "        edgecolor ='grey', label ='f1-score') \n",
    "   \n",
    "# Adding Xticks  \n",
    "plt.xlabel('Model', fontweight ='bold') \n",
    "plt.ylabel('Performance', fontweight ='bold') \n",
    "plt.xticks([r + barWidth for r in range(len(prec))], \n",
    "           ['LR', 'DC', 'SVM', 'RF', 'GB','XGB','ANN']) \n",
    "plt.legend()  \n",
    "plt.title(\"Metrics for SMOTE-generated Data for Class-1\")\n",
    "plt.show() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    " \n",
    " raw rf     1       0.97      0.22      0.36 \n",
    "            1       0.17      0.02      0.04 without hyper tuning\n",
    "smote rf   1       0.10      0.75      0.18  with hyper tuning \n",
    "prec = [.07,.15,.16,.23,.25,.11]  \n",
    "recall= [.77,.76,.08,.62,.71,.05]\n",
    "f1score = [.13,.25,.11,.33,.37,.07]\n",
    " Arushi   \n",
    "LR raw    0 0  0\n",
    "LR smote .07 .77 .13\n",
    "\n",
    "dc raw  .78 .63 .70\n",
    "dc smote .15 .76 .25\n",
    "\n",
    "SVM raw   .46 .25 .32 hyper tuned polynomial kernel\n",
    "SVM smote .16 .08 .11 hyper tuned polynomial kernel\n",
    "\n",
    "raw rf   .98 .47 .64  \n",
    "smote rf  .23 .62 .33\n",
    "\n",
    "GB raw   .73 .57 .64 \n",
    "GB smote .25 .71 .37\n",
    "\n",
    "ANN       .38 .21 .27\n",
    "ANN smote .11 .05 .07\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "from xgboost import XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,\n",
       "              importance_type='gain', interaction_constraints='',\n",
       "              learning_rate=0.300000012, max_delta_step=0, max_depth=6,\n",
       "              min_child_weight=1, missing=nan, monotone_constraints='()',\n",
       "              n_estimators=100, n_jobs=0, num_parallel_tree=1, random_state=0,\n",
       "              reg_alpha=0, reg_lambda=1, scale_pos_weight=1, subsample=1,\n",
       "              tree_method='exact', validate_parameters=1, verbosity=None)"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_xgboost = XGBClassifier()\n",
    "model_xgboost.fit(df_Train, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9995604158311997\n",
      "Testing Accuracy:  0.9978406391708055\n",
      "[[27641    15]\n",
      " [   45    85]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     27656\n",
      "           1       0.85      0.65      0.74       130\n",
      "\n",
      "    accuracy                           1.00     27786\n",
      "   macro avg       0.92      0.83      0.87     27786\n",
      "weighted avg       1.00      1.00      1.00     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred_xgboost = model_xgboost.predict(df_Test) \n",
    "print_eval(y_pred_xgboost,model_xgboost)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.9961594225252183\n",
      "Testing Accuracy:  0.9925142157921255\n",
      "[[27477   179]\n",
      " [   29   101]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.99      1.00     27656\n",
      "           1       0.36      0.78      0.49       130\n",
      "\n",
      "    accuracy                           0.99     27786\n",
      "   macro avg       0.68      0.89      0.74     27786\n",
      "weighted avg       1.00      0.99      0.99     27786\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "# smote\n",
    "\n",
    "# training and predictions : random forest \n",
    "model_xgboost_smote = XGBClassifier()\n",
    "model_xgboost_smote.fit(X_sm,y_sm)\n",
    "y_pred_xgboost_smote = model_xgboost_smote.predict(df_Test)\n",
    "\n",
    "print_eval(y_pred_xgboost_smote,model_xgboost_smote)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}