credit-draft.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.metrics import f1_score, accuracy_score,confusion_matrix,classification_report,roc_auc_score, RocCurveDisplay\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler,MinMaxScaler\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import StratifiedKFold,GridSearchCV\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Preprocessing:\n",
    "\n",
    "1. Fill the missing gender values with the most frequent values\n",
    "2. Fill the missing income values by taking the mean and filling it with the mean\n",
    "3. Filling the missing values in birthday count and converting it Age_in_years\n",
    "4. Drop the null values in type_occupation\n",
    "5. Outlier treatment - annual income, experience, family count\n",
    "6. Binary Encoding\n",
    "7. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1548 entries, 0 to 1547\n",
      "Data columns (total 15 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   Ind_ID           1548 non-null   int64  \n",
      " 1   GENDER           1541 non-null   object \n",
      " 2   Car_Owner        1548 non-null   object \n",
      " 3   Propert_Owner    1548 non-null   object \n",
      " 4   CHILDREN         1548 non-null   int64  \n",
      " 5   Annual_income    1525 non-null   float64\n",
      " 6   Type_Income      1548 non-null   object \n",
      " 7   EDUCATION        1548 non-null   object \n",
      " 8   Marital_status   1548 non-null   object \n",
      " 9   Housing_type     1548 non-null   object \n",
      " 10  Birthday_count   1526 non-null   float64\n",
      " 11  Employed_days    1548 non-null   int64  \n",
      " 12  Type_Occupation  1060 non-null   object \n",
      " 13  Family_Members   1548 non-null   int64  \n",
      " 14  Rejected         1548 non-null   int64  \n",
      "dtypes: float64(2), int64(5), object(8)\n",
      "memory usage: 181.5+ KB\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv(\"credit_card_dataset.csv\")\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "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>GENDER</th>\n",
       "      <th>Car_Owner</th>\n",
       "      <th>Propert_Owner</th>\n",
       "      <th>CHILDREN</th>\n",
       "      <th>Annual_income</th>\n",
       "      <th>Type_Income</th>\n",
       "      <th>EDUCATION</th>\n",
       "      <th>Marital_status</th>\n",
       "      <th>Housing_type</th>\n",
       "      <th>Birthday_count</th>\n",
       "      <th>Employed_days</th>\n",
       "      <th>Type_Occupation</th>\n",
       "      <th>Family_Members</th>\n",
       "      <th>Rejected</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>123</th>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>Pensioner</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-23286.0</td>\n",
       "      <td>365243</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>717</th>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>247500.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-11268.0</td>\n",
       "      <td>-2545</td>\n",
       "      <td>Laborers</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1341</th>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>State servant</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Single / not married</td>\n",
       "      <td>Municipal apartment</td>\n",
       "      <td>-10660.0</td>\n",
       "      <td>-1935</td>\n",
       "      <td>Core staff</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1485</th>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Separated</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-14760.0</td>\n",
       "      <td>-1315</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1319</th>\n",
       "      <td>F</td>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>180000.0</td>\n",
       "      <td>Pensioner</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Widow</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-20395.0</td>\n",
       "      <td>365243</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>388</th>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>Pensioner</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Widow</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-21126.0</td>\n",
       "      <td>365243</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>834</th>\n",
       "      <td>F</td>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>Commercial associate</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Civil marriage</td>\n",
       "      <td>With parents</td>\n",
       "      <td>-14228.0</td>\n",
       "      <td>-5330</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>728</th>\n",
       "      <td>F</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>1</td>\n",
       "      <td>270000.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Separated</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-14052.0</td>\n",
       "      <td>-140</td>\n",
       "      <td>Sales staff</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>573</th>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>40500.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-12729.0</td>\n",
       "      <td>-4936</td>\n",
       "      <td>Cleaning staff</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1126</th>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Single / not married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>-15323.0</td>\n",
       "      <td>-8290</td>\n",
       "      <td>Laborers</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     GENDER Car_Owner Propert_Owner  CHILDREN  Annual_income  \\\n",
       "123       M         N             Y         0       135000.0   \n",
       "717       F         N             Y         0       247500.0   \n",
       "1341      F         N             Y         0       135000.0   \n",
       "1485      F         N             Y         0       135000.0   \n",
       "1319      F         Y             N         0       180000.0   \n",
       "388       F         N             Y         0       202500.0   \n",
       "834       F         Y             N         0       202500.0   \n",
       "728       F         Y             Y         1       270000.0   \n",
       "573       F         N             Y         0        40500.0   \n",
       "1126      M         N             N         0       202500.0   \n",
       "\n",
       "               Type_Income                      EDUCATION  \\\n",
       "123              Pensioner  Secondary / secondary special   \n",
       "717                Working  Secondary / secondary special   \n",
       "1341         State servant               Higher education   \n",
       "1485               Working  Secondary / secondary special   \n",
       "1319             Pensioner  Secondary / secondary special   \n",
       "388              Pensioner  Secondary / secondary special   \n",
       "834   Commercial associate  Secondary / secondary special   \n",
       "728                Working               Higher education   \n",
       "573                Working  Secondary / secondary special   \n",
       "1126               Working  Secondary / secondary special   \n",
       "\n",
       "            Marital_status         Housing_type  Birthday_count  \\\n",
       "123                Married    House / apartment        -23286.0   \n",
       "717                Married    House / apartment        -11268.0   \n",
       "1341  Single / not married  Municipal apartment        -10660.0   \n",
       "1485             Separated    House / apartment        -14760.0   \n",
       "1319                 Widow    House / apartment        -20395.0   \n",
       "388                  Widow    House / apartment        -21126.0   \n",
       "834         Civil marriage         With parents        -14228.0   \n",
       "728              Separated    House / apartment        -14052.0   \n",
       "573                Married    House / apartment        -12729.0   \n",
       "1126  Single / not married    House / apartment        -15323.0   \n",
       "\n",
       "      Employed_days Type_Occupation  Family_Members  Rejected  \n",
       "123          365243             NaN               2         0  \n",
       "717           -2545        Laborers               2         0  \n",
       "1341          -1935      Core staff               1         1  \n",
       "1485          -1315             NaN               1         0  \n",
       "1319         365243             NaN               1         0  \n",
       "388          365243             NaN               1         0  \n",
       "834           -5330             NaN               2         0  \n",
       "728            -140     Sales staff               2         0  \n",
       "573           -4936  Cleaning staff               2         0  \n",
       "1126          -8290        Laborers               1         0  "
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#since we dont require the ind_id column\n",
    "\n",
    "df = df.drop(\"Ind_ID\", axis=1)\n",
    "df.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Duplicate rows:  162\n",
      "Null Values:  GENDER               7\n",
      "Car_Owner            0\n",
      "Propert_Owner        0\n",
      "CHILDREN             0\n",
      "Annual_income       23\n",
      "Type_Income          0\n",
      "EDUCATION            0\n",
      "Marital_status       0\n",
      "Housing_type         0\n",
      "Birthday_count      22\n",
      "Employed_days        0\n",
      "Type_Occupation    488\n",
      "Family_Members       0\n",
      "Rejected             0\n",
      "dtype: int64\n",
      "After deleting duplicate rows:  0\n"
     ]
    }
   ],
   "source": [
    "print(\"Duplicate rows: \", df.duplicated().sum())\n",
    "print(\"Null Values: \", df.isnull().sum())\n",
    "\n",
    "df = df.drop_duplicates()\n",
    "print(\"After deleting duplicate rows: \", df.duplicated().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "#replacing irrelevant values\n",
    "df['Marital_status'].unique()\n",
    "df.Marital_status= df.Marital_status.replace('Civil marriage','Married')\n",
    "df.Marital_status= df.Marital_status.replace('Widow','Separated')\n",
    "df.Marital_status= df.Marital_status.replace('Single / not married','Single')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "#creating a new column with age\n",
    "df['Age'] = np.abs((df['Birthday_count'])/365)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "#converting employed days to experience in YEARS\n",
    "df['Work_experience']= np.abs(df['Employed_days'])/365"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "F    879\n",
       "M    507\n",
       "Name: GENDER, dtype: int64"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#filling missing values in Gender by most populated values\n",
    "df['GENDER'].fillna(value=df['GENDER'].mode()[0], inplace=True)\n",
    "df['GENDER'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "#filling missing values in annual income by using the calculated mean of each type\n",
    "df['Annual_income'].fillna(df.groupby('Type_Income')['Annual_income'].transform('mean'), inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GENDER               0\n",
       "Car_Owner            0\n",
       "Propert_Owner        0\n",
       "CHILDREN             0\n",
       "Annual_income        0\n",
       "Type_Income          0\n",
       "EDUCATION            0\n",
       "Marital_status       0\n",
       "Housing_type         0\n",
       "Birthday_count      22\n",
       "Employed_days        0\n",
       "Type_Occupation    438\n",
       "Family_Members       0\n",
       "Rejected             0\n",
       "Age                 22\n",
       "Work_experience      0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GENDER               0\n",
       "Car_Owner            0\n",
       "Propert_Owner        0\n",
       "CHILDREN             0\n",
       "Annual_income        0\n",
       "Type_Income          0\n",
       "EDUCATION            0\n",
       "Marital_status       0\n",
       "Housing_type         0\n",
       "Birthday_count      22\n",
       "Employed_days        0\n",
       "Type_Occupation    438\n",
       "Family_Members       0\n",
       "Rejected             0\n",
       "Age                  0\n",
       "Work_experience      0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#filling missing values in age by calculating mean using the income type\n",
    "df['Age'].fillna(df.groupby('Type_Income')['Age'].transform('mean'), inplace=True)\n",
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "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>GENDER</th>\n",
       "      <th>Car_Owner</th>\n",
       "      <th>Propert_Owner</th>\n",
       "      <th>CHILDREN</th>\n",
       "      <th>Annual_income</th>\n",
       "      <th>Type_Income</th>\n",
       "      <th>EDUCATION</th>\n",
       "      <th>Marital_status</th>\n",
       "      <th>Housing_type</th>\n",
       "      <th>Family_Members</th>\n",
       "      <th>Rejected</th>\n",
       "      <th>Age</th>\n",
       "      <th>Work_experience</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>M</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>180000.0</td>\n",
       "      <td>Pensioner</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>51.430137</td>\n",
       "      <td>1000.665753</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>F</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>2</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>43.180822</td>\n",
       "      <td>8.693151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>247500.0</td>\n",
       "      <td>Commercial associate</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Separated</td>\n",
       "      <td>Rented apartment</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>46.619178</td>\n",
       "      <td>3.690411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>M</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>157500.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>27.197260</td>\n",
       "      <td>2.268493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>2</td>\n",
       "      <td>216000.0</td>\n",
       "      <td>State servant</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>42.312329</td>\n",
       "      <td>8.526027</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  GENDER Car_Owner Propert_Owner  CHILDREN  Annual_income  \\\n",
       "0      M         Y             Y         0       180000.0   \n",
       "1      F         Y             Y         2       135000.0   \n",
       "2      F         N             Y         0       247500.0   \n",
       "3      M         Y             Y         0       157500.0   \n",
       "4      F         N             Y         2       216000.0   \n",
       "\n",
       "            Type_Income                      EDUCATION Marital_status  \\\n",
       "0             Pensioner               Higher education        Married   \n",
       "1               Working  Secondary / secondary special        Married   \n",
       "2  Commercial associate               Higher education      Separated   \n",
       "3               Working  Secondary / secondary special        Married   \n",
       "4         State servant               Higher education        Married   \n",
       "\n",
       "        Housing_type  Family_Members  Rejected        Age  Work_experience  \n",
       "0  House / apartment               2         1  51.430137      1000.665753  \n",
       "1  House / apartment               4         0  43.180822         8.693151  \n",
       "2   Rented apartment               1         0  46.619178         3.690411  \n",
       "3  House / apartment               2         0  27.197260         2.268493  \n",
       "4  House / apartment               4         0  42.312329         8.526027  "
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#since there is alot of missing values in the type_occupation\n",
    "#instead of dropping the values (rows) we will be dropping the the whole column\n",
    "#along with the birthday counts and employed days that we already converted\n",
    "\n",
    "df.drop(columns = ['Type_Occupation', 'Birthday_count', 'Employed_days'] , inplace = True, axis = 1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "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>GENDER</th>\n",
       "      <th>Car_Owner</th>\n",
       "      <th>Propert_Owner</th>\n",
       "      <th>CHILDREN</th>\n",
       "      <th>Annual_income</th>\n",
       "      <th>Type_Income</th>\n",
       "      <th>EDUCATION</th>\n",
       "      <th>Marital_status</th>\n",
       "      <th>Housing_type</th>\n",
       "      <th>Family_Members</th>\n",
       "      <th>Rejected</th>\n",
       "      <th>Age</th>\n",
       "      <th>Work_experience</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>180000.0</td>\n",
       "      <td>Pensioner</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>51.430137</td>\n",
       "      <td>1000.665753</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>43.180822</td>\n",
       "      <td>8.693151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>247500.0</td>\n",
       "      <td>Commercial associate</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Separated</td>\n",
       "      <td>Rented apartment</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>46.619178</td>\n",
       "      <td>3.690411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>157500.0</td>\n",
       "      <td>Working</td>\n",
       "      <td>Secondary / secondary special</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>27.197260</td>\n",
       "      <td>2.268493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>216000.0</td>\n",
       "      <td>State servant</td>\n",
       "      <td>Higher education</td>\n",
       "      <td>Married</td>\n",
       "      <td>House / apartment</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>42.312329</td>\n",
       "      <td>8.526027</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   GENDER  Car_Owner  Propert_Owner  CHILDREN  Annual_income  \\\n",
       "0       1          1              1         0       180000.0   \n",
       "1       0          1              1         2       135000.0   \n",
       "2       0          0              1         0       247500.0   \n",
       "3       1          1              1         0       157500.0   \n",
       "4       0          0              1         2       216000.0   \n",
       "\n",
       "            Type_Income                      EDUCATION Marital_status  \\\n",
       "0             Pensioner               Higher education        Married   \n",
       "1               Working  Secondary / secondary special        Married   \n",
       "2  Commercial associate               Higher education      Separated   \n",
       "3               Working  Secondary / secondary special        Married   \n",
       "4         State servant               Higher education        Married   \n",
       "\n",
       "        Housing_type  Family_Members  Rejected        Age  Work_experience  \n",
       "0  House / apartment               2         1  51.430137      1000.665753  \n",
       "1  House / apartment               4         0  43.180822         8.693151  \n",
       "2   Rented apartment               1         0  46.619178         3.690411  \n",
       "3  House / apartment               2         0  27.197260         2.268493  \n",
       "4  House / apartment               4         0  42.312329         8.526027  "
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Binary Encoding\n",
    "df['GENDER'] = df['GENDER'].map({'M':1,'F':0})\n",
    "df['Car_Owner'] = df['Car_Owner'].map({'Y':1,'N':0})\n",
    "df['Propert_Owner'] = df['Propert_Owner'].map({'Y':1,'N':0})\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "#mapping education by ranks - 1, 2, 3, 4\n",
    "\n",
    "education_mapping = {\n",
    "    'Lower secondary': 0,\n",
    "    'Secondary / secondary special': 1,\n",
    "    'Incomplete higher': 2,\n",
    "    'Higher education': 3,\n",
    "    'Academic degree': 4\n",
    "}\n",
    "\n",
    "df['EDUCATION'] = df['EDUCATION'].map(education_mapping)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "#one hot encoding - converts categorical variables into a binary matrix of each category\n",
    "df = pd.get_dummies(df, columns=['Type_Income','Marital_status','Housing_type'],drop_first=True)  #drop_first to avoid multicolinearity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "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>GENDER</th>\n",
       "      <th>Car_Owner</th>\n",
       "      <th>Propert_Owner</th>\n",
       "      <th>CHILDREN</th>\n",
       "      <th>Annual_income</th>\n",
       "      <th>EDUCATION</th>\n",
       "      <th>Family_Members</th>\n",
       "      <th>Rejected</th>\n",
       "      <th>Age</th>\n",
       "      <th>Work_experience</th>\n",
       "      <th>Type_Income_Pensioner</th>\n",
       "      <th>Type_Income_State servant</th>\n",
       "      <th>Type_Income_Working</th>\n",
       "      <th>Marital_status_Separated</th>\n",
       "      <th>Marital_status_Single</th>\n",
       "      <th>Housing_type_House / apartment</th>\n",
       "      <th>Housing_type_Municipal apartment</th>\n",
       "      <th>Housing_type_Office apartment</th>\n",
       "      <th>Housing_type_Rented apartment</th>\n",
       "      <th>Housing_type_With parents</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>180000.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>51.430137</td>\n",
       "      <td>1000.665753</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>43.180822</td>\n",
       "      <td>8.693151</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>247500.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>46.619178</td>\n",
       "      <td>3.690411</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>157500.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>27.197260</td>\n",
       "      <td>2.268493</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>216000.0</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>42.312329</td>\n",
       "      <td>8.526027</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   GENDER  Car_Owner  Propert_Owner  CHILDREN  Annual_income  EDUCATION  \\\n",
       "0       1          1              1         0       180000.0          3   \n",
       "1       0          1              1         2       135000.0          1   \n",
       "2       0          0              1         0       247500.0          3   \n",
       "3       1          1              1         0       157500.0          1   \n",
       "4       0          0              1         2       216000.0          3   \n",
       "\n",
       "   Family_Members  Rejected        Age  Work_experience  \\\n",
       "0               2         1  51.430137      1000.665753   \n",
       "1               4         0  43.180822         8.693151   \n",
       "2               1         0  46.619178         3.690411   \n",
       "3               2         0  27.197260         2.268493   \n",
       "4               4         0  42.312329         8.526027   \n",
       "\n",
       "   Type_Income_Pensioner  Type_Income_State servant  Type_Income_Working  \\\n",
       "0                      1                          0                    0   \n",
       "1                      0                          0                    1   \n",
       "2                      0                          0                    0   \n",
       "3                      0                          0                    1   \n",
       "4                      0                          1                    0   \n",
       "\n",
       "   Marital_status_Separated  Marital_status_Single  \\\n",
       "0                         0                      0   \n",
       "1                         0                      0   \n",
       "2                         1                      0   \n",
       "3                         0                      0   \n",
       "4                         0                      0   \n",
       "\n",
       "   Housing_type_House / apartment  Housing_type_Municipal apartment  \\\n",
       "0                               1                                 0   \n",
       "1                               1                                 0   \n",
       "2                               0                                 0   \n",
       "3                               1                                 0   \n",
       "4                               1                                 0   \n",
       "\n",
       "   Housing_type_Office apartment  Housing_type_Rented apartment  \\\n",
       "0                              0                              0   \n",
       "1                              0                              0   \n",
       "2                              0                              1   \n",
       "3                              0                              0   \n",
       "4                              0                              0   \n",
       "\n",
       "   Housing_type_With parents  \n",
       "0                          0  \n",
       "1                          0  \n",
       "2                          0  \n",
       "3                          0  \n",
       "4                          0  "
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Annual Income before treating the outliers\n",
    "plt.subplot(1,2,1)\n",
    "plt.hist(df['Annual_income'], bins=30, color='blue', density=True, alpha=0.7)\n",
    "plt.axvline(x=np.mean(df['Annual_income']), color=\"red\", label=\"Mean\")\n",
    "plt.axvline(x=np.median(df['Annual_income']), color=\"green\", label=\"Median\")\n",
    "plt.grid()\n",
    "plt.legend()\n",
    "\n",
    "plt.subplot(1,2,2)  # nrows=1, ncols=2\n",
    "sns.boxplot(data=df, x='Annual_income')\n",
    "\n",
    "plt.subplots_adjust(right=2.0)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "def treating_outliers(col, df):\n",
    "    Q1 = np.percentile(df[col], 25 , interpolation='midpoint')\n",
    "    Q3 = np.percentile(df[col], 75 ,interpolation='midpoint')\n",
    "\n",
    "    IQR = Q3-Q1\n",
    "\n",
    "    upper_limit = Q3 + 1.5 * IQR\n",
    "    upper_limit\n",
    "    return df[df[col] < upper_limit]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = treating_outliers('Work_experience', df)\n",
    "df = treating_outliers('Annual_income', df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Annual income after treating the outliers\n",
    "plt.subplot(1,2,1)\n",
    "sns.histplot(data=df, x='Annual_income',kde=True)\n",
    "plt.axvline(x=df['Annual_income'].mean(),color=\"red\",label=\"Mean\")\n",
    "plt.axvline(x=df['Annual_income'].median(),color=\"green\",label=\"Median\")\n",
    "plt.grid()\n",
    "plt.legend()\n",
    "\n",
    "plt.subplot(1,2,2)  # nrows=1, ncols=2\n",
    "sns.boxplot(data=df, x='Annual_income')\n",
    "\n",
    "plt.subplots_adjust(right=2.0)\n",
    "plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#family members is a different case with one particular outlier\n",
    "plt.figure(figsize = (4,4))\n",
    "sns.boxplot(data = df , x = 'Family_Members')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[df['Family_Members'] < 6] #setting an upper limit manually"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU4AAAFzCAYAAABRt+AaAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAT3ElEQVR4nO3df2xV9f3H8ddpC7e/bqEQEDqurUOLv0oBWUzVOaUEKIqC4gTcYHZElwFxQ+OGGiHRiM7gL1yczoBoxIqQ6rIVQyMUtIuslBQZmkV0GmapVRfpbWcrbT/7wy/327u22He5vedano/kJj33x/l87qH36T331nM855wTAKDPkvyeAAB81xBOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMUvr7wM7OTtXX1ysYDMrzvFjOCQB84ZxTOBxWTk6OkpJ6f1/Z73DW19crFAr19+EAkLCOHDmicePG9Xp7v8MZDAYjA2RlZfV3NQCQMJqamhQKhSJ9602/w3li9zwrK4twAhhUvu3jR74cAgAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBACjFL8nAKmzs1PHjh2TJAUCAXme5/OMElNqairbBgmBcCaAY8eOad68eX5PI+Ft375daWlpfk8DYFcdAKx4x5lgmgtukBuS6vc0EobX2a7Mupf8ngYQhXAmGJeULCUP8XsaCcP5PQGgB+yqA4AR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoBRSjwHc86ptbVVkpSamirP8+I5PDBo8FryV1zfcba2tqqkpEQlJSWRf3QAdryW/MWuOgAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYEU4AMCKcAGBEOAHAiHACgBHhBAAjwgkARoQTAIwIJwAYpfg9AQCItSuuuCLyc1VVVczXzztOAIPKtm3bTrocC4QTwKCyfv36ky7HQlx31Z1zkZ9bW1vjOXRCi9oWrvf7nZb4nelR123R9XV1uluwYEGv15eVlcVsnD6Hs62tTW1tbZHlpqYm82BdHz9v3jzz408Lne1+zyCxdNke/M70rK2tTenp6X5Pw3fhcFgNDQ093tbQ0KBwOKxgMBiTsfq8q7527VoNGzYscgmFQjGZAADEwuLFi0/pdos+v+NctWqVVq5cGVluamoyxzMQCER+Li8vV2pqqunxg9WXX36phQsXfrOQxB86ROmyPfid+X+tra2Rd+BdX1ens+eff15z5sw56e2x0udXaSAQOOV/IM/zIj+npqYqLS3tlNY3WER9duf1fr/TEr8z36rr6+p0FgwGNWbMmB5318eOHRuz3XSJb9UBDCK9fQH00ksvxXQcwglgUFmxYsVJl2OBcAIYVK6//vqTLscC30QAGHQG4n+z7Ip3nABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGKXEc7DU1FRt37498jOA/uG15K+4htPzPKWlpcVzSGBQ4rXkL3bVAcCIcAKAEeEEACPCCQBGhBMAjAgnABgRTgAwIpwAYEQ4AcCIcAKAEeEEACPCCQBGhBMAjAgnABgRTgAwIpwAYEQ4AcCIcAKAEeEEACPCCQBGhBMAjAgnABgRTgAwIpwAYEQ4AcCIcAKAEeEEACPCCQBGhBMAjAgnABgRTgAwIpwAYEQ4AcCIcAKAEeEEACPCCQBGhBMAjAgnABgRTgAwIpwAYEQ4AcCIcAKAEeEEACPCCQBGhBMAjAgnABgRTgAwIpwAYEQ4AcCIcAKAEeEEACPCCQBGhBMAjAgnABgRTgAwIpwAYEQ4AcAoxe8JIJrX2SHXcdzvaSQMr7Pd7ykA3RDOBJN58BW/pwDgW7CrDgBGvONMAMOGDVN5ebkkKRAIyPM8n2eUmFJTU/2eAiCJcCaEpKQkZWdn+z0NAH3ErjoAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGBFOADAinABglNLfBzrnJElNTU0xmwwA+OlEz070rTf9Dmc4HJYkhUKh/q4CABJSOBzWsGHDer3dc9+W1l50dnaqvr5ewWBQnuf1+XFNTU0KhUI6cuSIsrKy+jP0oMR26R3bpmdsl971d9s45xQOh5WTk6OkpN4/yez3O86kpCSNGzeuvw9XVlYW/9g9YLv0jm3TM7ZL7/qzbU72TvMEvhwCACPCCQBGcQ9nIBDQ6tWrFQgE4j10QmO79I5t0zO2S+8Getv0+8shADhdsasOAEaEEwCMCCcAGBFOADCKWzj37NmjOXPmKCcnR57n6dVXX43X0Alt7dq1+sEPfqBgMKjRo0dr7ty5+sc//uH3tHz31FNPaeLEiZE/YC4qKtL27dv9nlZCWrt2rTzP069+9Su/p+KrNWvWyPO8qMuYMWMGZKy4hbOlpUWFhYV68skn4zXkd8Lu3bu1bNkyvf3226qsrFR7e7tmzJihlpYWv6fmq3HjxunBBx/Uvn37tG/fPk2bNk3XXnutDh065PfUEkpNTY2eeeYZTZw40e+pJIQLLrhAR48ejVwOHjw4MAM5H0hy5eXlfgyd8BobG50kt3v3br+nknCys7Pds88+6/c0EkY4HHbnnHOOq6ysdD/60Y/cbbfd5veUfLV69WpXWFgYl7H4jDPBHDt2TJI0YsQIn2eSODo6OlRWVqaWlhYVFRX5PZ2EsWzZMl111VWaPn2631NJGO+//75ycnJ01llnacGCBfrwww8HZJx+H+QDseec08qVK3XZZZfpwgsv9Hs6vjt48KCKiorU2tqqzMxMlZeX6/zzz/d7WgmhrKxM+/fvV01Njd9TSRgXX3yxnn/+eeXn5+vTTz/V/fffr0suuUSHDh3SyJEjYzoW4Uwgy5cv1zvvvKO33nrL76kkhAkTJqiurk5ffvmltm3bpiVLlmj37t2nfTyPHDmi2267TTt27FBqaqrf00kYJSUlkZ8LCgpUVFSk8ePHa9OmTVq5cmVMxyKcCWLFihX605/+pD179pzS4foGk6FDh+rss8+WJE2dOlU1NTV6/PHH9fTTT/s8M3/V1taqsbFRF110UeS6jo4O7dmzR08++aTa2tqUnJzs4wwTQ0ZGhgoKCvT+++/HfN2E02fOOa1YsULl5eWqqqrSWWed5feUEpZzTm1tbX5Pw3fFxcXdvi2++eabde655+o3v/kN0fw/bW1teu+99/TDH/4w5uuOWzibm5t1+PDhyPI///lP1dXVacSIETrzzDPjNY2Es2zZMm3evFmvvfaagsGgGhoaJH1zMNW0tDSfZ+efu+66SyUlJQqFQgqHwyorK1NVVZVef/11v6fmu2Aw2O0z8IyMDI0cOfK0/mz8jjvu0Jw5c3TmmWeqsbFR999/v5qamrRkyZLYDxaX7+6dc7t27XKSul2WLFkSrykkpJ62iSS3ceNGv6fmq9LSUpebm+uGDh3qRo0a5YqLi92OHTv8nlbC4s+RnLvxxhvd2LFj3ZAhQ1xOTo677rrr3KFDhwZkLA4rBwBG/B0nABgRTgAwIpwAYEQ4AcCIcAKAEeEEACPCCQBGhBNmeXl5euyxxyLLg/GI/v/7HIGuCOcg8LOf/azbKQM8z4v6X1xjqaamRrfccsuArPuKK66Q53l68MEHu902e/ZseZ6nNWvWDMjYQF8RzkFi1qxZUacMOHr06IAdMGTUqFFKT08fkHVLUigU0saNG6Ouq6+v186dOzV27NgBG3egHT9+3O8pIEYI5yARCAQ0ZsyYqMvjjz+ugoICZWRkKBQK6Ze//KWam5sjj3nuuec0fPhw/fnPf9aECROUnp6u+fPnq6WlRZs2bVJeXp6ys7O1YsUKdXR0RB53st3YadOmafny5VHXffHFFwoEAtq5c2efnsvVV1+tL774QtXV1VFznTFjhkaPHh1136+//lp33nmnvve97ykjI0MXX3yxqqqqTvk5SlI4HNaiRYuUmZmpnJwcrV+/Pur2Y8eO6ZZbbtHo0aOVlZWladOm6cCBA5Hb16xZo0mTJmnDhg36/ve/r0AgIOectm7dqoKCAqWlpWnkyJGaPn36aX+Oqe8awjmIJSUl6YknntDf//53bdq0STt37tSdd94ZdZ///Oc/euKJJ1RWVqbXX39dVVVVuu6661RRUaGKigq98MILeuaZZ7R169Y+jbl06VJt3rw56vBvL774onJycnTllVf2aR1Dhw7VTTfdFPWu87nnnlNpaWm3+958882qrq5WWVmZ3nnnHd1www2aNWtW1DEY+/scH374YU2cOFH79+/XqlWr9Otf/1qVlZWSvjnE3VVXXaWGhgZVVFSotrZWU6ZMUXFxsf79739H1nH48GFt2bJF27ZtU11dnRoaGrRw4UKVlpbqvffei8yFQ0Z8xwzIoUMQV0uWLHHJyckuIyMjcpk/f363+23ZssWNHDkysrxx40YnyR0+fDhy3a233urS09NdOByOXDdz5kx36623RpZzc3Pdo48+GllWl5Pvtba2uhEjRriXX345cvukSZPcmjVr+vRcThzl58CBAy4YDLrm5ma3e/duN3r0aPf111+7wsJCt3r1auecc4cPH3ae57lPPvkkah3FxcVu1apVp/wcZ82aFbXeG2+80ZWUlDjnnHvjjTdcVlaWa21tjbrP+PHj3dNPP+2c++bkYUOGDHGNjY2R22tra50k99FHH/VpeyAxcSDjQeLKK6/UU089FVnOyMjQrl279MADD+jdd99VU1OT2tvb1draqpaWFmVkZEiS0tPTNX78+MjjzjjjDOXl5SkzMzPqusbGxj7NIxAI6Cc/+Yk2bNigH//4x6qrq9OBAwfM37pPnDhR55xzjrZu3apdu3bppz/9qYYMGRJ1n/3798s5p/z8/Kjr29raos4x09/n+L8nhisqKop8RFFbW6vm5uZu57L56quv9MEHH0SWc3NzNWrUqMhyYWGhiouLVVBQoJkzZ2rGjBmaP3++srOz+7JZkCAI5yCRkZEROc2EJH388ceaPXu2fvGLX+i+++7TiBEj9NZbb+nnP/951JcU/xsjz/N6vK6zs7PPc1m6dKkmTZqkf/3rX9qwYYOKi4uVm5trfk6lpaX6/e9/r3fffVd/+9vfut3e2dmp5ORk1dbWdjvqedcoxvI5ep4XGXvs2LFRn6eeMHz48MjPJ/4DdUJycrIqKyv117/+VTt27ND69et19913a+/evRz9/zuEcA5S+/btU3t7u9atW6ekpG8+yt6yZUtcxi4oKNDUqVP1xz/+UZs3b+72pUpfLVq0SHfccYcKCwt7PEHb5MmT1dHRocbGxgE5PcLbb7/dbfncc8+VJE2ZMkUNDQ1KSUlRXl6eab2e5+nSSy/VpZdeqnvvvVe5ubkqLy+P+QnFMHAI5yA1fvx4tbe3a/369ZozZ46qq6v1hz/8IW7jL126VMuXL1d6errmzZvXr3VkZ2fr6NGj3d4dnpCfn6+bbrpJixcv1rp16zR58mR9/vnn2rlzpwoKCjR79uxTeQqqrq7W7373O82dO1eVlZV65ZVX9Je//EWSNH36dBUVFWnu3Ll66KGHNGHCBNXX16uiokJz587V1KlTe1zn3r179cYbb0T+QmDv3r367LPPdN55553SXBFffKs+SE2aNEmPPPKIHnroIV144YV68cUXtXbt2riNv3DhQqWkpGjRokWndArb4cOHd9vd7Wrjxo1avHixbr/9dk2YMEHXXHON9u7dq1Ao1O8xT7j99ttVW1uryZMn67777tO6des0c+ZMSd+8a6yoqNDll1+u0tJS5efna8GCBfroo490xhln9LrOrKws7dmzR7Nnz1Z+fr7uuecerVu3LurUtkh8nDoDA+LIkSPKy8tTTU2NpkyZ4vd0gJginIip48eP6+jRo/rtb3+rjz/+OOqP2IHBgl11xFR1dbVyc3NVW1vb7TPVN998U5mZmb1egO8K3nEibr766it98sknvd7e9c+pgERGOAHAiF11ADAinABgRDgBwIhwAoAR4QQAI8IJAEaEEwCMCCcAGP0X5ZpBeT09pLUAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (4,4))\n",
    "sns.boxplot(data = df , x = 'Family_Members')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#we have a highly imbalanced data\n",
    "plt.figure(figsize = (6, 6))\n",
    "sns.countplot(data = df, x = 'Rejected')\n",
    "plt.title(\"Distribution of Output variable)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    967\n",
       "1     93\n",
       "Name: Rejected, dtype: int64"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Rejected'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "def performing_smote(X, Y, k=5, oversample_ratio=1.0):\n",
    "    \"\"\"\n",
    "    Custom implementation of SMOTE oversampling.\n",
    "\n",
    "    Parameters:\n",
    "    - X: Feature matrix\n",
    "    - Y: Target variable\n",
    "    - k: Number of nearest neighbors to consider\n",
    "    - oversample_ratio: Ratio of oversampling (e.g., 1.0 means balance the classes)\n",
    "\n",
    "    Returns:\n",
    "    - Oversampled feature matrix and target variable as DataFrames\n",
    "    \"\"\"\n",
    "    # Separate minority and majority classes\n",
    "    minority_class = X[Y == 1]\n",
    "    majority_class = X[Y == 0]\n",
    "\n",
    "    # Calculate the number of samples to generate\n",
    "    num_samples = int(len(minority_class) * oversample_ratio)\n",
    "\n",
    "    # Randomly select samples from the minority class\n",
    "    random_indices = np.random.choice(len(minority_class), num_samples, replace=True)\n",
    "    minority_samples = minority_class[random_indices]\n",
    "\n",
    "    # Compute Euclidean distances between minority samples\n",
    "    distances = np.linalg.norm(minority_samples[:, np.newaxis, :] - minority_class, axis=2)\n",
    "\n",
    "    # Find k-nearest neighbors\n",
    "    nearest_neighbors_indices = np.argpartition(distances, k, axis=1)[:, :k]\n",
    "\n",
    "    # Generate synthetic samples\n",
    "    synthetic_samples = []\n",
    "    for i, neighbors in enumerate(nearest_neighbors_indices):\n",
    "        for _ in range(int(oversample_ratio)):\n",
    "            neighbor = minority_class[neighbors[np.random.randint(k)]]\n",
    "            alpha = np.random.rand()\n",
    "            synthetic_sample = minority_samples[i] + alpha * (neighbor - minority_samples[i])\n",
    "            synthetic_samples.append(synthetic_sample)\n",
    "\n",
    "    # Combine minority samples, majority samples, and synthetic samples\n",
    "    oversampled_X = pd.DataFrame(np.vstack((X, np.array(synthetic_samples))))\n",
    "    oversampled_Y = pd.DataFrame(np.hstack((Y, np.ones(len(synthetic_samples)))))\n",
    "\n",
    "    return oversampled_X, oversampled_Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [],
   "source": [
    "def oversample_smote_df(X, y, k_neighbors=5):\n",
    "    # Separate data into majority and minority classes\n",
    "    majority_indices = np.where(y == 0)[0]\n",
    "    minority_indices = np.where(y == 1)[0]\n",
    "\n",
    "    # Calculate the number of synthetic samples to generate\n",
    "    num_synthetic_samples = len(majority_indices) - len(minority_indices)\n",
    "\n",
    "    # Randomly select k neighbors for each sample in the minority class\n",
    "    synthetic_samples = []\n",
    "    for minority_index in minority_indices:\n",
    "        neighbors_indices = np.random.choice(minority_indices, size=k_neighbors, replace=True)\n",
    "        neighbors = X[neighbors_indices]\n",
    "\n",
    "        # Create synthetic samples by combining the minority sample with its neighbors\n",
    "        for _ in range(num_synthetic_samples // len(minority_indices)):\n",
    "            weights = np.random.uniform(0, 1, size=X.shape[1])\n",
    "            synthetic_sample = X[minority_index] + weights.mean() * (neighbors.mean(axis=0) - X[minority_index])\n",
    "            synthetic_samples.append(synthetic_sample)\n",
    "\n",
    "    # Combine the original majority class with the synthetic samples to create the oversampled arrays\n",
    "    x_oversampled = np.vstack([X, np.array(synthetic_samples)])\n",
    "    y_oversampled = np.concatenate([y, np.ones(len(synthetic_samples), dtype=int)])\n",
    "\n",
    "    # Create DataFrames for the oversampled data\n",
    "    columns = df.columns.drop('Rejected')\n",
    "    X_df = pd.DataFrame(x_oversampled, columns=columns)\n",
    "    y_df = pd.DataFrame(y_oversampled, columns=['Rejected'])\n",
    "\n",
    "    return X_df, y_df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pairwise_distances(from_data, to_data):\n",
    "    return np.sqrt(np.sum((from_data[:, np.newaxis, :] - to_data[np.newaxis, :, :])**2,axis=2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {},
   "outputs": [],
   "source": [
    "def custom_smote(X, y, k=5, oversample_ratio=1.0): #no3\n",
    "    # Find minority class samples\n",
    "    minority_indices = np.where(y == 1)[0]\n",
    "    minority_samples = X[minority_indices]\n",
    "\n",
    "    # Calculate pairwise distances\n",
    "    distances = pairwise_distances(minority_samples, minority_samples)\n",
    "\n",
    "    # Generate synthetic samples\n",
    "    synthetic_samples = []\n",
    "    for i, sample in enumerate(minority_samples):\n",
    "        nearest_neighbors = np.argsort(distances[i])[1:k+1]  # Exclude the sample itself\n",
    "\n",
    "        for neighbor_index in nearest_neighbors:\n",
    "            alpha = np.random.uniform(0, 1, size=X.shape[1])\n",
    "            synthetic_sample = sample + alpha * (minority_samples[neighbor_index] - sample)\n",
    "            synthetic_samples.append(synthetic_sample)\n",
    "\n",
    "    # Combine original and synthetic samples\n",
    "    synthetic_samples = np.array(synthetic_samples)\n",
    "    # X_resampled = np.vstack([X, synthetic_samples])\n",
    "    # y_resampled = np.hstack([y, np.ones(len(synthetic_samples))])\n",
    "\n",
    "    oversampled_X = pd.DataFrame(np.vstack((X, np.array(synthetic_samples))))\n",
    "    oversampled_Y = pd.DataFrame(np.hstack((Y, np.ones(len(synthetic_samples)))))\n",
    "\n",
    "    return oversampled_X, oversampled_Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_resampled, y_resampled = custom_smote(X, Y, k=5, oversample_ratio=1.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop(columns=['Rejected']).to_numpy()\n",
    "Y = df['Rejected'].to_numpy()\n",
    "XOS, YOS = oversample_smote_df(X, Y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop(columns=['Rejected']).to_numpy()\n",
    "Y = df['Rejected'].to_numpy()\n",
    "\n",
    "Xos, Yos = performing_smote(X, Y, k=5, oversample_ratio=1.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    967\n",
       "1.0    558\n",
       "dtype: int64"
      ]
     },
     "execution_count": 251,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_resampled.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    967\n",
       "1.0    186\n",
       "dtype: int64"
      ]
     },
     "execution_count": 181,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Yos.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "def manual_train_test_split(X, Y, test_size=0.2, random_state=None):\n",
    "    if random_state is not None:\n",
    "        np.random.seed(random_state)\n",
    "\n",
    "    # Shuffle the indices\n",
    "    indices = np.arange(len(X))\n",
    "    np.random.shuffle(indices)\n",
    "\n",
    "    # Calculate the split index\n",
    "    split_index = int((1 - test_size) * len(X))\n",
    "    if split_index <= 0 or split_index >= len(X):\n",
    "        raise ValueError(\"Invalid split index. Check the size of your dataset.\")\n",
    "\n",
    "    # Split the data\n",
    "    X_train, X_test = X[indices[:split_index]], X[indices[split_index:]]\n",
    "    y_train, y_test = y[indices[:split_index]], y[indices[split_index:]]\n",
    "\n",
    "    return X_train, X_test, Y_train, Y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyError",
     "evalue": "'[1108, 665, 421, 246, 1459, 286, 1460, 1374, 672, 69, 403, 45, 1081, 349, 365, 1352, 405, 532, 734, 363, 333, 1301, 501, 1484, 1184, 1230, 901, 1494, 678, 268, 1209, 1283, 181, 1187, 76, 183, 1097, 776, 263, 398, 861, 1504, 1039, 476, 745, 1292, 142, 1164, 1242, 199, 137, 1062, 1133, 239, 248, 287, 1501, 715, 550, 1299, 952, 903, 197, 624, 835, 652, 1322, 503, 512, 1388, 781, 1123, 1312, 1150, 494, 1139, 564, 498, 145, 921, 472, 157, 204, 1437, 1036, 411, 928, 1507, 1421, 547, 691, 436, 1171, 1440, 675, 1411, 1303, 834, 1428, 827, 272, 71, 1327, 883, 1136, 968, 143, 1470, 68, 58, 278, 481, 1182, 948, 504, 1446, 1339, 925, 922, 1477, 724, 782, 757, 659, 418, 736, 488, 1202, 590, 696, 1229, 1147, 453, 931, 1331, 101, 1134, 941, 1037, 1452, 814, 1464, 542, 1104, 728, 756, 65, 1056, 1353, 508, 1224, 880, 487, 318, 557, 1176, 1244, 455, 1120, 1509, 792, 705, 579, 553, 1505, 1228, 606, 370, 46, 1250, 1321, 452, 1285, 1488, 1152, 549, 226, 1391, 798, 771, 1449, 173, 556, 844, 1241, 1183, 1173, 273, 1487, 1252, 1365, 1082, 929, 540, 305, 1445, 1028, 230, 114, 332, 123, 1130, 701, 1094, 796, 1181, 308, 1376, 479, 484, 1430, 640, 535, 1194, 1044, 593, 1054, 33, 336, 598, 502, 1412, 1342, 369, 419, 874, 819, 1455, 803, 52, 1245, 356, 663, 1007, 1492, 1268, 241, 1266, 799, 1014, 763, 93, 1026, 1240, 1091, 575, 462, 464, 280, 161, 1008, 1191, 1316, 1220, 112, 1156, 109, 649, 735, 188, 1317, 1226, 385, 1486, 307, 681, 352, 387, 526, 1306, 704, 690, 932, 794, 1160, 1290, 311, 283, 457, 256, 212, 259, 584, 492, 870, 235, 367, 1348, 448, 1387, 21, 1434, 544, 826, 694, 1247, 70, 1367, 788, 682, 1003, 20, 600, 720, 1121, 1217, 1330, 1105, 525, 128, 1523, 84, 1211, 926, 612, 341, 323, 1061, 1500, 773, 1328, 881, 664, 1095, 944, 857, 1090, 1511, 431, 102, 1260, 994, 1372, 47, 354, 185, 171, 400, 732, 859, 824, 61, 807, 737, 1414, 667, 285, 194, 963, 845, 951, 1233, 571, 1467, 1163, 1271, 1221, 653, 1357, 417, 1256, 611, 662, 871, 127, 402, 361, 743, 169, 1493, 119, 993, 996, 1297, 1142, 1524, 1496, 1407, 470, 200, 1310, 249, 377, 315, 146, 465, 744, 1400, 1448, 141, 711, 1167, 729, 355, 1046, 335, 67, 1489, 1481, 1058, 82, 325, 1213, 1023, 158, 1148, 863, 1456, 618, 560, 1154, 1158, 1196, 1518, 739, 1341, 496, 603, 846, 99, 328, 468, 1329, 83, 1178, 1480, 1084, 1280, 1122, 582, 215, 1264, 477, 1476, 80, 1018, 37, 1432, 505, 541, 619, 391, 182, 374, 519, 1001, 608, 66, 995, 717, 1205, 407, 873, 1416, 1186, 172, 580, 1279, 330, 810, 284, 978, 1249, 1197, 1239, 1258, 232, 1510, 568, 748, 1067, 625, 1089, 930, 572, 627, 954, 721, 1499, 1124, 651, 270, 614, 460, 573, 688, 1227, 60, 201, 265, 789, 699, 399, 1027, 279, 301, 795, 915, 1131, 28, 129, 240, 1513, 666, 888, 1273, 916, 529, 451, 866, 126, 961, 347, 1116, 116, 104, 829, 379, 351, 293, 1236, 887, 1272, 1308, 91, 23, 1042, 1088, 90, 1040, 523, 1307, 658, 953, 1113, 956, 1006, 1442, 1072, 779, 983, 601, 645, 223, 1096, 213, 825, 1325, 1140, 848, 77, 1092, 388, 597, 439, 997, 843, 586, 821, 1254, 955, 551, 1415, 297, 545, 543, 277, 1405, 322, 490, 1420, 802, 1168, 980, 966, 1395, 1144, 1015, 957, 1198, 221, 1385, 1334, 837, 1417, 1345, 309, 1463, 1053, 281, 1276, 1323, 148, 397, 851, 327, 486, 577, 677, 522, 1201, 1031, 271, 643, 1146, 86, 164, 1315, 842, 555, 193, 326, 343, 108, 155, 530, 1216, 375, 738, 1389, 770, 480, 1392, 1064, 940, 1443, 1288, 471, 1284, 765, 695, 75, 134, 437, 1403, 319, 991, 914, 1318, 1426, 321, 567, 897, 251, 1212, 1098, 1361, 62, 1351, 89, 1427, 719, 156, 441, 1274, 731, 414, 1101, 552, 511, 1401, 521, 909, 1267, 708, 473, 1025, 761, 1473, 180, 727, 755, 1378, 401, 1516, 813, 345, 886, 250, 295, 1170, 981, 853, 1383, 262, 120, 1359, 1246, 219, 267, 368, 192, 290, 692, 697, 790, 1390, 1450, 877, 546, 74, 1368, 784, 252, 432, 1153, 596, 1424, 768, 654, 988, 999, 209, 787, 801, 1366, 179, 1048, 513, 312, 440, 621, 160, 911, 376, 205, 288, 56, 684, 384, 500, 1137, 949, 231, 1275, 425, 1218, 741, 310, 589, 237, 236, 426, 967, 1356, 581, 1135, 175, 1141, 1203, 1114, 1471, 456, 816, 1491, 823, 1466, 317, 1522, 982, 671, 1289, 152, 558, 986, 1382, 923, 1248, 1286, 1380, 752, 786, 1363, 574, 1320, 264, 291, 1305, 599, 977, 154, 1043, 22, 862, 1429, 1502, 885, 1059, 1419, 809, 258, 227, 316, 946, 107, 1394, 1166, 427, 85, 1185, 1118, 1077, 592, 610, 1506, 524, 30, 960, 438, 342, 710, 1336, 458, 1257, 669, 1132, 125, 257, 1199, 847, 313, 891, 1065, 1068, 198, 628, 409, 1281, 716, 849, 174, 1282, 1086, 839, 469, 202, 971, 595, 631, 177, 275, 907, 412, 483, 1468, 1457, 475, 1051, 538, 615, 984, 117, 536, 1127, 1298, 641, 602, 1210, 1189, 1034, 163, 111, 344, 1190, 1479, 482, 1332, 1277, 130, 1293, 1300, 338, 905, 1358, 72, 514, 969, 515, 1222, 378, 447, 985, 208, 303, 576, 620, 943, 1066, 565, 1381, 693, 1291, 1373, 644, 882, 340, 917, 346, 1438, 1324, 304, 36, 811, 42, 1237, 394, 878, 1313, 1223, 733, 459, 707, 423, 1497, 945, 879, 294, 585, 26, 680, 1055, 1423, 135, 510, 210, 1115, 353, 964, 1259, 852, 55, 463, 1433, 1454, 856, 216, 1371, 40, 884, 396, 1465, 1063, 1251, 1396, 998, 1408, 178, 176, 740, 507, 570, 920, 1333, 300, 1435, 990, 639, 563, 1263, 78, 518, 1078, 726, 228, 818, 1076, 50, 29, 633, 656, 561, 1346, 840, 225, 1143, 203, 444, 371, 32, 1129, 1215, 965, 1393, 893, 668, 1214, 1238, 958, 334, 382, 1145, 1278, 372, 151, 959, 778, 222, 386, 702, 642, 429, 1195, 1355, 1304, 445, 854, 1458, 41, 833, 113, 48, 289, 1021, 206, 973, 632, 775, 889, 899, 832, 683, 1379, 617, 685, 869, 673, 749, 1002, 1079, 408, 629, 1398, 646, 139, 1125, 913, 165, 1100, 812, 214, 131, 314, 138, 1200, 416, 1029, 38, 53, 706, 136, 1384, 793, 35, 25, 657, 449, 578, 302, 1422, 1475, 1165, 1052, 1193, 777, 1503, 244, 103, 276, 393, 110, 588, 159, 970, 747, 1179, 1111, 660, 1161, 339, 329, 133, 637, 167, 1397, 927, 860, 1020, 1482, 24, 806, 1413, 443, 150, 1309, 94, 1354, 217, 381, 1047, 780, 900, 1436, 864, 292, 712, 100, 1049, 783, 1206, 489, 700, 1155, 750, 962, 1012, 242, 168, 73, 559, 1319, 1495, 1270, 118, 1138, 520, 713, 992, 254, 1093, 1038, 383, 1102, 516, 54, 467, 266, 1033, 583, 766, 537, 1441, 59, 1177, 166, 1265, 815, 331, 247, 373, 1204, 499, 1207, 1180, 636, 616, 1370, 1174, 566, 234, 626, 430, 1010, 1364, 191, 1188, 1175, 435, 836, 1255, 1016, 184, 687, 638, 569, 933, 1169, 533, 1520, 1517, 742, 196, 1515, 495, 591, 936, 1041, 979, 380, 27, 395, 320, 147, 797, 623, 238, 1009, 910, 895, 1335, 106, 350, 650, 389, 1004, 1117, 554, 1439, 170, 1253, 306, 415, 1485, 1296, 428, 186, 1005, 908, 269, 867, 630, 950, 162, 413, 1344, 95, 98, 1030, 876, 817, 822, 207, 974, 686, 872, 245, 229, 906, 896, 1418, 1402, 1447, 865, 1338, 791, 1022, 189, 260, 144, 828, 698, 919, 1410, 937, 1035, 1377, 890, 594, 531, 894, 1483, 1151, 57, 1057, 1261, 282, 934, 362, 243, 92, 1234, 337, 1013, 947, 1109, 190, 39, 975, 634, 764, 723, 1128, 63, 233, 676, 105, 753, 902, 253, 422, 774, 1232, 858, 924, 79, 989, 1106, 497, 1462, 87, 64, 1294, 605, 274, 1512, 1362, 1444, 153, 613, 410, 830, 912, 517, 44, 390, 220, 140, 357, 722, 1157, 261, 1302, 898, 1103, 195, 115, 1075, 1073, 718, 1472, 1208, 1369, 358, 1060, 1019, 1000, 935, 754, 420, 149, 1069, 1119, 1451, 534, 548, 298, 976, 1314, 1337, 122, 1399, 1360, 1159, 1406] not in index'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 37\u001b[0m line \u001b[0;36m1\n\u001b[0;32m----> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X50sZmlsZQ%3D%3D?line=0'>1</a>\u001b[0m X_train, X_test, Y_train, Y_test \u001b[39m=\u001b[39m manual_train_test_split(X_resampled, y_resampled, test_size\u001b[39m=\u001b[39m\u001b[39m0.1\u001b[39m, random_state\u001b[39m=\u001b[39m\u001b[39m2\u001b[39m)\n",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 37\u001b[0m line \u001b[0;36m1\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X50sZmlsZQ%3D%3D?line=11'>12</a>\u001b[0m     \u001b[39mraise\u001b[39;00m \u001b[39mValueError\u001b[39;00m(\u001b[39m\"\u001b[39m\u001b[39mInvalid split index. Check the size of your dataset.\u001b[39m\u001b[39m\"\u001b[39m)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X50sZmlsZQ%3D%3D?line=13'>14</a>\u001b[0m \u001b[39m# Split the data\u001b[39;00m\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X50sZmlsZQ%3D%3D?line=14'>15</a>\u001b[0m X_train, X_test \u001b[39m=\u001b[39m X[indices[:split_index]], X[indices[split_index:]]\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X50sZmlsZQ%3D%3D?line=15'>16</a>\u001b[0m y_train, y_test \u001b[39m=\u001b[39m y[indices[:split_index]], y[indices[split_index:]]\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X50sZmlsZQ%3D%3D?line=17'>18</a>\u001b[0m \u001b[39mreturn\u001b[39;00m X_train, X_test, Y_train, Y_test\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/pandas/core/frame.py:3813\u001b[0m, in \u001b[0;36mDataFrame.__getitem__\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   3811\u001b[0m     \u001b[39mif\u001b[39;00m is_iterator(key):\n\u001b[1;32m   3812\u001b[0m         key \u001b[39m=\u001b[39m \u001b[39mlist\u001b[39m(key)\n\u001b[0;32m-> 3813\u001b[0m     indexer \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mcolumns\u001b[39m.\u001b[39m_get_indexer_strict(key, \u001b[39m\"\u001b[39m\u001b[39mcolumns\u001b[39m\u001b[39m\"\u001b[39m)[\u001b[39m1\u001b[39m]\n\u001b[1;32m   3815\u001b[0m \u001b[39m# take() does not accept boolean indexers\u001b[39;00m\n\u001b[1;32m   3816\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mgetattr\u001b[39m(indexer, \u001b[39m\"\u001b[39m\u001b[39mdtype\u001b[39m\u001b[39m\"\u001b[39m, \u001b[39mNone\u001b[39;00m) \u001b[39m==\u001b[39m \u001b[39mbool\u001b[39m:\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/pandas/core/indexes/base.py:6070\u001b[0m, in \u001b[0;36mIndex._get_indexer_strict\u001b[0;34m(self, key, axis_name)\u001b[0m\n\u001b[1;32m   6067\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[1;32m   6068\u001b[0m     keyarr, indexer, new_indexer \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_reindex_non_unique(keyarr)\n\u001b[0;32m-> 6070\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_raise_if_missing(keyarr, indexer, axis_name)\n\u001b[1;32m   6072\u001b[0m keyarr \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mtake(indexer)\n\u001b[1;32m   6073\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39misinstance\u001b[39m(key, Index):\n\u001b[1;32m   6074\u001b[0m     \u001b[39m# GH 42790 - Preserve name from an Index\u001b[39;00m\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/pandas/core/indexes/base.py:6133\u001b[0m, in \u001b[0;36mIndex._raise_if_missing\u001b[0;34m(self, key, indexer, axis_name)\u001b[0m\n\u001b[1;32m   6130\u001b[0m     \u001b[39mraise\u001b[39;00m \u001b[39mKeyError\u001b[39;00m(\u001b[39mf\u001b[39m\u001b[39m\"\u001b[39m\u001b[39mNone of [\u001b[39m\u001b[39m{\u001b[39;00mkey\u001b[39m}\u001b[39;00m\u001b[39m] are in the [\u001b[39m\u001b[39m{\u001b[39;00maxis_name\u001b[39m}\u001b[39;00m\u001b[39m]\u001b[39m\u001b[39m\"\u001b[39m)\n\u001b[1;32m   6132\u001b[0m not_found \u001b[39m=\u001b[39m \u001b[39mlist\u001b[39m(ensure_index(key)[missing_mask\u001b[39m.\u001b[39mnonzero()[\u001b[39m0\u001b[39m]]\u001b[39m.\u001b[39munique())\n\u001b[0;32m-> 6133\u001b[0m \u001b[39mraise\u001b[39;00m \u001b[39mKeyError\u001b[39;00m(\u001b[39mf\u001b[39m\u001b[39m\"\u001b[39m\u001b[39m{\u001b[39;00mnot_found\u001b[39m}\u001b[39;00m\u001b[39m not in index\u001b[39m\u001b[39m\"\u001b[39m)\n",
      "\u001b[0;31mKeyError\u001b[0m: '[1108, 665, 421, 246, 1459, 286, 1460, 1374, 672, 69, 403, 45, 1081, 349, 365, 1352, 405, 532, 734, 363, 333, 1301, 501, 1484, 1184, 1230, 901, 1494, 678, 268, 1209, 1283, 181, 1187, 76, 183, 1097, 776, 263, 398, 861, 1504, 1039, 476, 745, 1292, 142, 1164, 1242, 199, 137, 1062, 1133, 239, 248, 287, 1501, 715, 550, 1299, 952, 903, 197, 624, 835, 652, 1322, 503, 512, 1388, 781, 1123, 1312, 1150, 494, 1139, 564, 498, 145, 921, 472, 157, 204, 1437, 1036, 411, 928, 1507, 1421, 547, 691, 436, 1171, 1440, 675, 1411, 1303, 834, 1428, 827, 272, 71, 1327, 883, 1136, 968, 143, 1470, 68, 58, 278, 481, 1182, 948, 504, 1446, 1339, 925, 922, 1477, 724, 782, 757, 659, 418, 736, 488, 1202, 590, 696, 1229, 1147, 453, 931, 1331, 101, 1134, 941, 1037, 1452, 814, 1464, 542, 1104, 728, 756, 65, 1056, 1353, 508, 1224, 880, 487, 318, 557, 1176, 1244, 455, 1120, 1509, 792, 705, 579, 553, 1505, 1228, 606, 370, 46, 1250, 1321, 452, 1285, 1488, 1152, 549, 226, 1391, 798, 771, 1449, 173, 556, 844, 1241, 1183, 1173, 273, 1487, 1252, 1365, 1082, 929, 540, 305, 1445, 1028, 230, 114, 332, 123, 1130, 701, 1094, 796, 1181, 308, 1376, 479, 484, 1430, 640, 535, 1194, 1044, 593, 1054, 33, 336, 598, 502, 1412, 1342, 369, 419, 874, 819, 1455, 803, 52, 1245, 356, 663, 1007, 1492, 1268, 241, 1266, 799, 1014, 763, 93, 1026, 1240, 1091, 575, 462, 464, 280, 161, 1008, 1191, 1316, 1220, 112, 1156, 109, 649, 735, 188, 1317, 1226, 385, 1486, 307, 681, 352, 387, 526, 1306, 704, 690, 932, 794, 1160, 1290, 311, 283, 457, 256, 212, 259, 584, 492, 870, 235, 367, 1348, 448, 1387, 21, 1434, 544, 826, 694, 1247, 70, 1367, 788, 682, 1003, 20, 600, 720, 1121, 1217, 1330, 1105, 525, 128, 1523, 84, 1211, 926, 612, 341, 323, 1061, 1500, 773, 1328, 881, 664, 1095, 944, 857, 1090, 1511, 431, 102, 1260, 994, 1372, 47, 354, 185, 171, 400, 732, 859, 824, 61, 807, 737, 1414, 667, 285, 194, 963, 845, 951, 1233, 571, 1467, 1163, 1271, 1221, 653, 1357, 417, 1256, 611, 662, 871, 127, 402, 361, 743, 169, 1493, 119, 993, 996, 1297, 1142, 1524, 1496, 1407, 470, 200, 1310, 249, 377, 315, 146, 465, 744, 1400, 1448, 141, 711, 1167, 729, 355, 1046, 335, 67, 1489, 1481, 1058, 82, 325, 1213, 1023, 158, 1148, 863, 1456, 618, 560, 1154, 1158, 1196, 1518, 739, 1341, 496, 603, 846, 99, 328, 468, 1329, 83, 1178, 1480, 1084, 1280, 1122, 582, 215, 1264, 477, 1476, 80, 1018, 37, 1432, 505, 541, 619, 391, 182, 374, 519, 1001, 608, 66, 995, 717, 1205, 407, 873, 1416, 1186, 172, 580, 1279, 330, 810, 284, 978, 1249, 1197, 1239, 1258, 232, 1510, 568, 748, 1067, 625, 1089, 930, 572, 627, 954, 721, 1499, 1124, 651, 270, 614, 460, 573, 688, 1227, 60, 201, 265, 789, 699, 399, 1027, 279, 301, 795, 915, 1131, 28, 129, 240, 1513, 666, 888, 1273, 916, 529, 451, 866, 126, 961, 347, 1116, 116, 104, 829, 379, 351, 293, 1236, 887, 1272, 1308, 91, 23, 1042, 1088, 90, 1040, 523, 1307, 658, 953, 1113, 956, 1006, 1442, 1072, 779, 983, 601, 645, 223, 1096, 213, 825, 1325, 1140, 848, 77, 1092, 388, 597, 439, 997, 843, 586, 821, 1254, 955, 551, 1415, 297, 545, 543, 277, 1405, 322, 490, 1420, 802, 1168, 980, 966, 1395, 1144, 1015, 957, 1198, 221, 1385, 1334, 837, 1417, 1345, 309, 1463, 1053, 281, 1276, 1323, 148, 397, 851, 327, 486, 577, 677, 522, 1201, 1031, 271, 643, 1146, 86, 164, 1315, 842, 555, 193, 326, 343, 108, 155, 530, 1216, 375, 738, 1389, 770, 480, 1392, 1064, 940, 1443, 1288, 471, 1284, 765, 695, 75, 134, 437, 1403, 319, 991, 914, 1318, 1426, 321, 567, 897, 251, 1212, 1098, 1361, 62, 1351, 89, 1427, 719, 156, 441, 1274, 731, 414, 1101, 552, 511, 1401, 521, 909, 1267, 708, 473, 1025, 761, 1473, 180, 727, 755, 1378, 401, 1516, 813, 345, 886, 250, 295, 1170, 981, 853, 1383, 262, 120, 1359, 1246, 219, 267, 368, 192, 290, 692, 697, 790, 1390, 1450, 877, 546, 74, 1368, 784, 252, 432, 1153, 596, 1424, 768, 654, 988, 999, 209, 787, 801, 1366, 179, 1048, 513, 312, 440, 621, 160, 911, 376, 205, 288, 56, 684, 384, 500, 1137, 949, 231, 1275, 425, 1218, 741, 310, 589, 237, 236, 426, 967, 1356, 581, 1135, 175, 1141, 1203, 1114, 1471, 456, 816, 1491, 823, 1466, 317, 1522, 982, 671, 1289, 152, 558, 986, 1382, 923, 1248, 1286, 1380, 752, 786, 1363, 574, 1320, 264, 291, 1305, 599, 977, 154, 1043, 22, 862, 1429, 1502, 885, 1059, 1419, 809, 258, 227, 316, 946, 107, 1394, 1166, 427, 85, 1185, 1118, 1077, 592, 610, 1506, 524, 30, 960, 438, 342, 710, 1336, 458, 1257, 669, 1132, 125, 257, 1199, 847, 313, 891, 1065, 1068, 198, 628, 409, 1281, 716, 849, 174, 1282, 1086, 839, 469, 202, 971, 595, 631, 177, 275, 907, 412, 483, 1468, 1457, 475, 1051, 538, 615, 984, 117, 536, 1127, 1298, 641, 602, 1210, 1189, 1034, 163, 111, 344, 1190, 1479, 482, 1332, 1277, 130, 1293, 1300, 338, 905, 1358, 72, 514, 969, 515, 1222, 378, 447, 985, 208, 303, 576, 620, 943, 1066, 565, 1381, 693, 1291, 1373, 644, 882, 340, 917, 346, 1438, 1324, 304, 36, 811, 42, 1237, 394, 878, 1313, 1223, 733, 459, 707, 423, 1497, 945, 879, 294, 585, 26, 680, 1055, 1423, 135, 510, 210, 1115, 353, 964, 1259, 852, 55, 463, 1433, 1454, 856, 216, 1371, 40, 884, 396, 1465, 1063, 1251, 1396, 998, 1408, 178, 176, 740, 507, 570, 920, 1333, 300, 1435, 990, 639, 563, 1263, 78, 518, 1078, 726, 228, 818, 1076, 50, 29, 633, 656, 561, 1346, 840, 225, 1143, 203, 444, 371, 32, 1129, 1215, 965, 1393, 893, 668, 1214, 1238, 958, 334, 382, 1145, 1278, 372, 151, 959, 778, 222, 386, 702, 642, 429, 1195, 1355, 1304, 445, 854, 1458, 41, 833, 113, 48, 289, 1021, 206, 973, 632, 775, 889, 899, 832, 683, 1379, 617, 685, 869, 673, 749, 1002, 1079, 408, 629, 1398, 646, 139, 1125, 913, 165, 1100, 812, 214, 131, 314, 138, 1200, 416, 1029, 38, 53, 706, 136, 1384, 793, 35, 25, 657, 449, 578, 302, 1422, 1475, 1165, 1052, 1193, 777, 1503, 244, 103, 276, 393, 110, 588, 159, 970, 747, 1179, 1111, 660, 1161, 339, 329, 133, 637, 167, 1397, 927, 860, 1020, 1482, 24, 806, 1413, 443, 150, 1309, 94, 1354, 217, 381, 1047, 780, 900, 1436, 864, 292, 712, 100, 1049, 783, 1206, 489, 700, 1155, 750, 962, 1012, 242, 168, 73, 559, 1319, 1495, 1270, 118, 1138, 520, 713, 992, 254, 1093, 1038, 383, 1102, 516, 54, 467, 266, 1033, 583, 766, 537, 1441, 59, 1177, 166, 1265, 815, 331, 247, 373, 1204, 499, 1207, 1180, 636, 616, 1370, 1174, 566, 234, 626, 430, 1010, 1364, 191, 1188, 1175, 435, 836, 1255, 1016, 184, 687, 638, 569, 933, 1169, 533, 1520, 1517, 742, 196, 1515, 495, 591, 936, 1041, 979, 380, 27, 395, 320, 147, 797, 623, 238, 1009, 910, 895, 1335, 106, 350, 650, 389, 1004, 1117, 554, 1439, 170, 1253, 306, 415, 1485, 1296, 428, 186, 1005, 908, 269, 867, 630, 950, 162, 413, 1344, 95, 98, 1030, 876, 817, 822, 207, 974, 686, 872, 245, 229, 906, 896, 1418, 1402, 1447, 865, 1338, 791, 1022, 189, 260, 144, 828, 698, 919, 1410, 937, 1035, 1377, 890, 594, 531, 894, 1483, 1151, 57, 1057, 1261, 282, 934, 362, 243, 92, 1234, 337, 1013, 947, 1109, 190, 39, 975, 634, 764, 723, 1128, 63, 233, 676, 105, 753, 902, 253, 422, 774, 1232, 858, 924, 79, 989, 1106, 497, 1462, 87, 64, 1294, 605, 274, 1512, 1362, 1444, 153, 613, 410, 830, 912, 517, 44, 390, 220, 140, 357, 722, 1157, 261, 1302, 898, 1103, 195, 115, 1075, 1073, 718, 1472, 1208, 1369, 358, 1060, 1019, 1000, 935, 754, 420, 149, 1069, 1119, 1451, 534, 548, 298, 976, 1314, 1337, 122, 1399, 1360, 1159, 1406] not in index'"
     ]
    }
   ],
   "source": [
    "X_train, X_test, Y_train, Y_test = manual_train_test_split(X_resampled, y_resampled, test_size=0.1, random_state=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_test_split_custom(*arrays, test_size=None, random_state=None):\n",
    "    n_samples = len(arrays[0])\n",
    "    test_size = 0.2 if test_size is None else test_size\n",
    "\n",
    "    if random_state:\n",
    "        np.random.seed(random_state)\n",
    "\n",
    "    num_test_samples = int(test_size * n_samples)\n",
    "    indices = np.random.permutation(n_samples)\n",
    "    test_indices = indices[:num_test_samples]\n",
    "    train_indices = indices[num_test_samples:]\n",
    "\n",
    "    # Create the training and testing sets\n",
    "    result = tuple([array.iloc[train_indices] for array in arrays] + [array.iloc[test_indices] for array in arrays])\n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_trainn, X_testt, y_trainn, y_testt = train_test_split_custom(X_resampled, y_resampled, test_size=0.20, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random Forest Classifier Attempt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_split(index, value, X, y):\n",
    "    left_mask = X[:, index] < value\n",
    "    right_mask = ~left_mask\n",
    "    return X[left_mask], X[right_mask], y[left_mask], y[right_mask]\n",
    "\n",
    "# Function to calculate Gini index\n",
    "def gini_index(groups, classes):\n",
    "    total_samples = sum(len(group) for group in groups)\n",
    "    gini = 0.0\n",
    "    for group in groups:\n",
    "        group_size = len(group)\n",
    "        if group_size == 0:\n",
    "            continue\n",
    "        proportion = group_size / total_samples\n",
    "        gini += (1.0 - np.sum((np.sum(group == cls) / group_size) ** 2 for cls in classes)) * proportion\n",
    "    return gini\n",
    "\n",
    "# Function to find the best split point for a dataset\n",
    "# def get_split(X, y, n_features):\n",
    "#     classes = np.unique(y)\n",
    "#     b_index, b_value, b_score, b_groups = 999, 999, 999, None\n",
    "#     features = np.random.choice(X.shape[1], n_features, replace=False)\n",
    "#     for index in features:\n",
    "#         for row in range(X.shape[0]):\n",
    "#             groups = test_split(index, X[row, index], X, y)\n",
    "#             gini = gini_index(groups, classes)\n",
    "#             if gini < b_score:\n",
    "#                 b_index, b_value, b_score, b_groups = index, X[row, index], gini, groups\n",
    "#     return {'index': b_index, 'value': b_value, 'groups': b_groups}\n",
    "\n",
    "def get_split(X, y, n_features):\n",
    "    classes = np.unique(y)\n",
    "    b_index, b_value, b_score, b_groups = 999, 999, 999, None\n",
    "    features = np.random.choice(X.shape[1], n_features, replace=False)\n",
    "    for index in features:\n",
    "        for row in range(X.shape[0]):\n",
    "            groups = test_split(index, X[row, index], X, y)\n",
    "            if groups is None:\n",
    "                continue\n",
    "                \n",
    "            gini = gini_index(groups, classes)\n",
    "            if gini < b_score:\n",
    "                b_index, b_value, b_score, b_groups = index, X[row, index], gini, groups\n",
    "    return {'index': b_index, 'value': b_value, 'groups': b_groups}\n",
    "# Function to create a terminal node value\n",
    "def to_terminal(group):\n",
    "    return np.argmax(np.bincount(group))\n",
    "\n",
    "# Recursive function to build a decision tree\n",
    "# def build_tree(X, y, max_depth, min_size, n_features, depth):\n",
    "#     node = get_split(X, y, n_features)\n",
    "#     left, right, left_y, right_y = node['groups']\n",
    "#     del node['groups']\n",
    "    \n",
    "#     if not left_y.tolist() or not right_y.tolist():\n",
    "#         node['left'] = node['right'] = to_terminal(left_y.tolist() + right_y.tolist())\n",
    "#         return node\n",
    "    \n",
    "#     if depth >= max_depth:\n",
    "#         node['left'], node['right'] = to_terminal(left_y), to_terminal(right_y)\n",
    "#         return node\n",
    "    \n",
    "#     node['left'] = build_tree(left, left_y, max_depth, min_size, n_features, depth + 1)\n",
    "#     node['right'] = build_tree(right, right_y, max_depth, min_size, n_features, depth + 1)\n",
    "    \n",
    "#     return node\n",
    "\n",
    "def build_tree(X, y, max_depth, min_size, n_features, depth):\n",
    "    node = get_split(X, y, n_features)\n",
    "    \n",
    "    if node['groups'] is None:\n",
    "        node['left'] = node['right'] = to_terminal(y.values.tolist())\n",
    "        return node\n",
    "    \n",
    "    left, right, left_y, right_y = node['groups']\n",
    "    del node['groups']\n",
    "    \n",
    "    if not left_y.tolist() or not right_y.tolist():\n",
    "        node['left'] = node['right'] = to_terminal(left_y.tolist() + right_y.tolist())\n",
    "        return node\n",
    "    \n",
    "    if depth >= max_depth:\n",
    "        node['left'], node['right'] = to_terminal(left_y), to_terminal(right_y)\n",
    "        return node\n",
    "    \n",
    "    node['left'] = build_tree(left, left_y, max_depth, min_size, n_features, depth + 1)\n",
    "    node['right'] = build_tree(right, right_y, max_depth, min_size, n_features, depth + 1)\n",
    "    \n",
    "    return node\n",
    "\n",
    "# Function to make a prediction with a decision tree\n",
    "def predict(node, row):\n",
    "    if row[node['index']] < node['value']:\n",
    "        if isinstance(node['left'], dict):\n",
    "            return predict(node['left'], row)\n",
    "        else:\n",
    "            return node['left']\n",
    "    else:\n",
    "        if isinstance(node['right'], dict):\n",
    "            return predict(node['right'], row)\n",
    "        else:\n",
    "            return node['right']\n",
    "\n",
    "# Function to create a random forest\n",
    "def random_forest(X, y, n_trees, max_depth, min_size, sample_size, n_features):\n",
    "    trees = []\n",
    "    for _ in range(n_trees):\n",
    "        sample_indices = np.random.choice(X.index, size=int(sample_size * X.shape[0]), replace=True)\n",
    "        \n",
    "        # Ensure sampled indices are within the range of both X and y\n",
    "        valid_indices = np.intersect1d(sample_indices, y.index)\n",
    "        sample_X, sample_y = X.loc[valid_indices], y.loc[valid_indices]\n",
    "        \n",
    "        tree = build_tree(sample_X, sample_y, max_depth, min_size, n_features, 1)\n",
    "        trees.append(tree)\n",
    "    return trees\n",
    "\n",
    "\n",
    "# Function to make a prediction with a random forest\n",
    "def random_forest_predict(trees, row):\n",
    "    predictions = [predict(tree, row) for tree in trees]\n",
    "    return max(set(predictions), key=predictions.count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "attempt to get argmax of an empty sequence",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 42\u001b[0m line \u001b[0;36m8\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=4'>5</a>\u001b[0m n_features \u001b[39m=\u001b[39m \u001b[39mint\u001b[39m(np\u001b[39m.\u001b[39msqrt(X_trainn\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m]))  \u001b[39m# Use square root of the number of features\u001b[39;00m\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=6'>7</a>\u001b[0m \u001b[39m# Build the random forest\u001b[39;00m\n\u001b[0;32m----> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=7'>8</a>\u001b[0m forest \u001b[39m=\u001b[39m random_forest(X_trainn, y_trainn, n_trees, max_depth, min_size, sample_size, n_features)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=9'>10</a>\u001b[0m \u001b[39m# Make predictions on the test set\u001b[39;00m\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=10'>11</a>\u001b[0m predictions \u001b[39m=\u001b[39m [random_forest_predict(forest, row) \u001b[39mfor\u001b[39;00m row \u001b[39min\u001b[39;00m X_testt]\n",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 42\u001b[0m line \u001b[0;36m1\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=110'>111</a>\u001b[0m     valid_indices \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mintersect1d(sample_indices, y\u001b[39m.\u001b[39mindex)\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=111'>112</a>\u001b[0m     sample_X, sample_y \u001b[39m=\u001b[39m X\u001b[39m.\u001b[39mloc[valid_indices], y\u001b[39m.\u001b[39mloc[valid_indices]\n\u001b[0;32m--> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=113'>114</a>\u001b[0m     tree \u001b[39m=\u001b[39m build_tree(sample_X, sample_y, max_depth, min_size, n_features, \u001b[39m1\u001b[39m)\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=114'>115</a>\u001b[0m     trees\u001b[39m.\u001b[39mappend(tree)\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=115'>116</a>\u001b[0m \u001b[39mreturn\u001b[39;00m trees\n",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 42\u001b[0m line \u001b[0;36m7\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=68'>69</a>\u001b[0m node \u001b[39m=\u001b[39m get_split(X, y, n_features)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=70'>71</a>\u001b[0m \u001b[39mif\u001b[39;00m node[\u001b[39m'\u001b[39m\u001b[39mgroups\u001b[39m\u001b[39m'\u001b[39m] \u001b[39mis\u001b[39;00m \u001b[39mNone\u001b[39;00m:\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=71'>72</a>\u001b[0m     node[\u001b[39m'\u001b[39m\u001b[39mleft\u001b[39m\u001b[39m'\u001b[39m] \u001b[39m=\u001b[39m node[\u001b[39m'\u001b[39m\u001b[39mright\u001b[39m\u001b[39m'\u001b[39m] \u001b[39m=\u001b[39m to_terminal(y\u001b[39m.\u001b[39mvalues\u001b[39m.\u001b[39mtolist())\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=72'>73</a>\u001b[0m     \u001b[39mreturn\u001b[39;00m node\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=74'>75</a>\u001b[0m left, right, left_y, right_y \u001b[39m=\u001b[39m node[\u001b[39m'\u001b[39m\u001b[39mgroups\u001b[39m\u001b[39m'\u001b[39m]\n",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 42\u001b[0m line \u001b[0;36m4\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=45'>46</a>\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39mto_terminal\u001b[39m(group):\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X52sZmlsZQ%3D%3D?line=46'>47</a>\u001b[0m     \u001b[39mreturn\u001b[39;00m np\u001b[39m.\u001b[39margmax(np\u001b[39m.\u001b[39mbincount(group))\n",
      "File \u001b[0;32m<__array_function__ internals>:200\u001b[0m, in \u001b[0;36margmax\u001b[0;34m(*args, **kwargs)\u001b[0m\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/numpy/core/fromnumeric.py:1242\u001b[0m, in \u001b[0;36margmax\u001b[0;34m(a, axis, out, keepdims)\u001b[0m\n\u001b[1;32m   1155\u001b[0m \u001b[39m\u001b[39m\u001b[39m\"\"\"\u001b[39;00m\n\u001b[1;32m   1156\u001b[0m \u001b[39mReturns the indices of the maximum values along an axis.\u001b[39;00m\n\u001b[1;32m   1157\u001b[0m \n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m   1239\u001b[0m \u001b[39m(2, 1, 4)\u001b[39;00m\n\u001b[1;32m   1240\u001b[0m \u001b[39m\"\"\"\u001b[39;00m\n\u001b[1;32m   1241\u001b[0m kwds \u001b[39m=\u001b[39m {\u001b[39m'\u001b[39m\u001b[39mkeepdims\u001b[39m\u001b[39m'\u001b[39m: keepdims} \u001b[39mif\u001b[39;00m keepdims \u001b[39mis\u001b[39;00m \u001b[39mnot\u001b[39;00m np\u001b[39m.\u001b[39m_NoValue \u001b[39melse\u001b[39;00m {}\n\u001b[0;32m-> 1242\u001b[0m \u001b[39mreturn\u001b[39;00m _wrapfunc(a, \u001b[39m'\u001b[39m\u001b[39margmax\u001b[39m\u001b[39m'\u001b[39m, axis\u001b[39m=\u001b[39maxis, out\u001b[39m=\u001b[39mout, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwds)\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/numpy/core/fromnumeric.py:57\u001b[0m, in \u001b[0;36m_wrapfunc\u001b[0;34m(obj, method, *args, **kwds)\u001b[0m\n\u001b[1;32m     54\u001b[0m     \u001b[39mreturn\u001b[39;00m _wrapit(obj, method, \u001b[39m*\u001b[39margs, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwds)\n\u001b[1;32m     56\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[0;32m---> 57\u001b[0m     \u001b[39mreturn\u001b[39;00m bound(\u001b[39m*\u001b[39margs, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwds)\n\u001b[1;32m     58\u001b[0m \u001b[39mexcept\u001b[39;00m \u001b[39mTypeError\u001b[39;00m:\n\u001b[1;32m     59\u001b[0m     \u001b[39m# A TypeError occurs if the object does have such a method in its\u001b[39;00m\n\u001b[1;32m     60\u001b[0m     \u001b[39m# class, but its signature is not identical to that of NumPy's. This\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     64\u001b[0m     \u001b[39m# Call _wrapit from within the except clause to ensure a potential\u001b[39;00m\n\u001b[1;32m     65\u001b[0m     \u001b[39m# exception has a traceback chain.\u001b[39;00m\n\u001b[1;32m     66\u001b[0m     \u001b[39mreturn\u001b[39;00m _wrapit(obj, method, \u001b[39m*\u001b[39margs, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwds)\n",
      "\u001b[0;31mValueError\u001b[0m: attempt to get argmax of an empty sequence"
     ]
    }
   ],
   "source": [
    "n_trees = 10\n",
    "max_depth = 5\n",
    "min_size = 1\n",
    "sample_size = 0.8  # Proportion of samples to use for each tree\n",
    "n_features = int(np.sqrt(X_trainn.shape[1]))  # Use square root of the number of features\n",
    "\n",
    "# Build the random forest\n",
    "forest = random_forest(X_trainn, y_trainn, n_trees, max_depth, min_size, sample_size, n_features)\n",
    "\n",
    "# Make predictions on the test set\n",
    "predictions = [random_forest_predict(forest, row) for row in X_testt]\n",
    "\n",
    "# Evaluate the accuracy\n",
    "# accuracy = np.sum(predictions == Y_test) / len(Y_test)\n",
    "# print(f\"Random Forest Accuracy: {accuracy:.4f}\")\n",
    "\n",
    "accuracy = np.sum(predictions == y_testt) / len(y_testt)\n",
    "print(f\"Random Forest Accuracy: {accuracy:.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_split(index, value, X, y):\n",
    "    left_mask = X[:, index] < value\n",
    "    right_mask = ~left_mask\n",
    "    return X[left_mask], X[right_mask], y[left_mask], y[right_mask]\n",
    "\n",
    "# Function to calculate Gini index\n",
    "def gini_index(groups, classes):\n",
    "    total_samples = sum(len(group) for group in groups)\n",
    "    gini = 0.0\n",
    "    for group in groups:\n",
    "        group_size = len(group)\n",
    "        if group_size == 0:\n",
    "            continue\n",
    "        proportion = group_size / total_samples\n",
    "        gini += (1.0 - np.sum((np.sum(group == cls) / group_size) ** 2 for cls in classes)) * proportion\n",
    "    return gini\n",
    "\n",
    "# Function to find the best split point for a dataset\n",
    "def get_split(X, y, n_features):\n",
    "    classes = np.unique(y)\n",
    "    b_index, b_value, b_score, b_groups = 999, 999, 999, None\n",
    "    features = np.random.choice(X.shape[1], n_features, replace=False)\n",
    "    for index in features:\n",
    "        for row in range(X.shape[0]):\n",
    "            groups = test_split(index, X[row, index], X, y)\n",
    "            gini = gini_index(groups, classes)\n",
    "            if gini < b_score:\n",
    "                b_index, b_value, b_score, b_groups = index, X[row, index], gini, groups\n",
    "    return {'index': b_index, 'value': b_value, 'groups': b_groups}\n",
    "\n",
    "# Function to create a terminal node value\n",
    "def to_terminal(group):\n",
    "    return np.argmax(np.bincount(group))\n",
    "\n",
    "# Recursive function to build a decision tree\n",
    "def build_tree(X, y, max_depth, min_size, n_features, depth):\n",
    "    node = get_split(X, y, n_features)\n",
    "    left, right, left_y, right_y = node['groups']\n",
    "    del node['groups']\n",
    "    \n",
    "    if not left_y.tolist() or not right_y.tolist():\n",
    "        node['left'] = node['right'] = to_terminal(left_y.tolist() + right_y.tolist())\n",
    "        return node\n",
    "    \n",
    "    if depth >= max_depth:\n",
    "        node['left'], node['right'] = to_terminal(left_y), to_terminal(right_y)\n",
    "        return node\n",
    "    \n",
    "    node['left'] = build_tree(left, left_y, max_depth, min_size, n_features, depth + 1)\n",
    "    node['right'] = build_tree(right, right_y, max_depth, min_size, n_features, depth + 1)\n",
    "    \n",
    "    return node\n",
    "\n",
    "# Function to make a prediction with a decision tree\n",
    "def predict(node, row):\n",
    "    if row[node['index']] < node['value']:\n",
    "        if isinstance(node['left'], dict):\n",
    "            return predict(node['left'], row)\n",
    "        else:\n",
    "            return node['left']\n",
    "    else:\n",
    "        if isinstance(node['right'], dict):\n",
    "            return predict(node['right'], row)\n",
    "        else:\n",
    "            return node['right']\n",
    "\n",
    "# Function to create a random forest\n",
    "def random_forest(X, y, n_trees, max_depth, min_size, sample_size, n_features):\n",
    "    trees = []\n",
    "    for _ in range(n_trees):\n",
    "        sample_indices = np.random.choice(X.shape[0], size=int(sample_size * X.shape[0]), replace=True)\n",
    "        sample_X, sample_y = X[sample_indices], y[sample_indices]\n",
    "        tree = build_tree(sample_X, sample_y, max_depth, min_size, n_features, 1)\n",
    "        trees.append(tree)\n",
    "    return trees\n",
    "\n",
    "# Function to make a prediction with a random forest\n",
    "def random_forest_predict(trees, row):\n",
    "    predictions = [predict(tree, row) for tree in trees]\n",
    "    return max(set(predictions), key=predictions.count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'numpy.ndarray' object has no attribute 'index'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 44\u001b[0m line \u001b[0;36m1\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=4'>5</a>\u001b[0m n_features \u001b[39m=\u001b[39m \u001b[39mint\u001b[39m(np\u001b[39m.\u001b[39msqrt(X_train\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m]))  \u001b[39m# Use square root of the number of features\u001b[39;00m\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=6'>7</a>\u001b[0m \u001b[39m# Build the random forest\u001b[39;00m\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=7'>8</a>\u001b[0m \u001b[39m# forest = random_forest(X_trainn, y_trainn, n_trees, max_depth, min_size, sample_size, n_features)\u001b[39;00m\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=9'>10</a>\u001b[0m forest \u001b[39m=\u001b[39m random_forest(X_train, Y_train, n_trees, max_depth, min_size, sample_size, n_features)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=11'>12</a>\u001b[0m \u001b[39m# Make predictions on the test set\u001b[39;00m\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=12'>13</a>\u001b[0m predictions \u001b[39m=\u001b[39m [random_forest_predict(forest, row) \u001b[39mfor\u001b[39;00m row \u001b[39min\u001b[39;00m X_test_scaled_new]\n",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 44\u001b[0m line \u001b[0;36m1\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=107'>108</a>\u001b[0m sample_indices \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mrandom\u001b[39m.\u001b[39mchoice(X\u001b[39m.\u001b[39mindex, size\u001b[39m=\u001b[39m\u001b[39mint\u001b[39m(sample_size \u001b[39m*\u001b[39m X\u001b[39m.\u001b[39mshape[\u001b[39m0\u001b[39m]), replace\u001b[39m=\u001b[39m\u001b[39mTrue\u001b[39;00m)\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=109'>110</a>\u001b[0m \u001b[39m# Ensure sampled indices are within the range of both X and y\u001b[39;00m\n\u001b[0;32m--> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=110'>111</a>\u001b[0m valid_indices \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mintersect1d(sample_indices, y\u001b[39m.\u001b[39mindex)\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=111'>112</a>\u001b[0m sample_X, sample_y \u001b[39m=\u001b[39m X\u001b[39m.\u001b[39mloc[valid_indices], y\u001b[39m.\u001b[39mloc[valid_indices]\n\u001b[1;32m    <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y142sZmlsZQ%3D%3D?line=113'>114</a>\u001b[0m tree \u001b[39m=\u001b[39m build_tree(sample_X, sample_y, max_depth, min_size, n_features, \u001b[39m1\u001b[39m)\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'numpy.ndarray' object has no attribute 'index'"
     ]
    }
   ],
   "source": [
    "n_trees = 10\n",
    "max_depth = 5\n",
    "min_size = 1\n",
    "sample_size = 0.8  # Proportion of samples to use for each tree\n",
    "n_features = int(np.sqrt(X_train.shape[1]))  # Use square root of the number of features\n",
    "\n",
    "# Build the random forest\n",
    "# forest = random_forest(X_trainn, y_trainn, n_trees, max_depth, min_size, sample_size, n_features)\n",
    "\n",
    "forest = random_forest(X_train, Y_train, n_trees, max_depth, min_size, sample_size, n_features)\n",
    "\n",
    "# Make predictions on the test set\n",
    "predictions = [random_forest_predict(forest, row) for row in X_test_scaled_new]\n",
    "\n",
    "# Evaluate the accuracy\n",
    "accuracy = np.sum(predictions == Y_test) / len(Y_test)\n",
    "print(f\"Random Forest Accuracy: {accuracy:.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "#feature scaling\n",
    "def min_max_scaling(data):\n",
    "    min_val = np.min(data, axis=0)\n",
    "    max_val = np.max(data, axis=0)\n",
    "    scaled_data = (data - min_val) / (max_val - min_val)\n",
    "    return scaled_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Training Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 1. 1. ... 0. 0. 0.]]\n",
      "\n",
      "Min-Max Scaled Training Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 1. 1. ... 0. 0. 0.]]\n",
      "\n",
      "Original Test Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 1. 0.]\n",
      " [1. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 1. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]]\n",
      "\n",
      "Min-Max Scaled Test Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 1. 0.]\n",
      " [1. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 1. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]]\n",
      "no error\n"
     ]
    }
   ],
   "source": [
    "if np.isnan(X_test).any():\n",
    "    # Handle missing values (e.g., by filling with mean or using another imputation strategy)\n",
    "    X_test = np.nan_to_num(X_test, nan=np.nanmean(X_test, axis=0))\n",
    "\n",
    "# Perform min-max scaling\n",
    "X_train_scaled = min_max_scaling(X_train)\n",
    "X_test_scaledd = min_max_scaling(X_test)\n",
    "\n",
    "print(\"Original Training Data:\")\n",
    "print(X_train)\n",
    "print(\"\\nMin-Max Scaled Training Data:\")\n",
    "print(X_train_scaled)\n",
    "\n",
    "print(\"\\nOriginal Test Data:\")\n",
    "print(X_test)\n",
    "print(\"\\nMin-Max Scaled Test Data:\")\n",
    "print(X_test_scaledd)\n",
    "\n",
    "if np.isnan(X_test).any():\n",
    "    print('error')\n",
    "else:\n",
    "    print('no error')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Training Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 1. 1. ... 0. 0. 0.]]\n",
      "\n",
      "Min-Max Scaled Training Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 1. 1. ... 0. 0. 0.]]\n",
      "\n",
      "Original Test Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 1. 0.]\n",
      " [1. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 1. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]]\n",
      "\n",
      "Min-Max Scaled Test Data:\n",
      "[[0. 0. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 1. 0.]\n",
      " [1. 0. 1. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 1. 1. ... 0. 0. 0.]\n",
      " [1. 1. 1. ... 0. 0. 0.]\n",
      " [0. 0. 1. ... 0. 0. 0.]]\n"
     ]
    }
   ],
   "source": [
    "X_train_scaled = min_max_scaling(X_train)\n",
    "X_test_scaled = min_max_scaling(X_test)\n",
    "\n",
    "print(\"Original Training Data:\")\n",
    "print(X_train)\n",
    "print(\"\\nMin-Max Scaled Training Data:\")\n",
    "print(X_train_scaled)\n",
    "\n",
    "print(\"\\nOriginal Test Data:\")\n",
    "print(X_test)\n",
    "print(\"\\nMin-Max Scaled Test Data:\")\n",
    "print(X_test_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train samples -> 848\n",
      "test samples -> 212\n"
     ]
    }
   ],
   "source": [
    "print('train samples ->', X_train.shape[0])\n",
    "print('test samples ->', X_test.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 257,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_size = 848\n",
    "test_size = 848\n",
    "X_trainX = X_resampled[:train_size]\n",
    "y_trainX = y_resampled[:train_size]\n",
    "\n",
    "X_testX = X_resampled[train_size:train_size + test_size]\n",
    "y_testX = y_resampled[train_size:train_size + test_size]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomMinMaxScaler:\n",
    "    def __init__(self):\n",
    "        self.min_val = None\n",
    "        self.max_val = None\n",
    "\n",
    "    def fit_transform(self, data):\n",
    "        self.min_val = np.min(data, axis=0)\n",
    "        self.max_val = np.max(data, axis=0)\n",
    "        scaled_data = (data - self.min_val) / (self.max_val - self.min_val)\n",
    "        return scaled_data\n",
    "\n",
    "    def transform(self, data):\n",
    "        scaled_data = (data - self.min_val) / (self.max_val - self.min_val)\n",
    "        return scaled_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Training Data:\n",
      "            0         1         2         3              4         5   \\\n",
      "788   0.000000  0.000000  1.000000  1.000000  157500.000000  1.000000   \n",
      "803   1.000000  1.000000  1.000000  2.000000  180000.000000  3.000000   \n",
      "1302  0.662759  0.000000  1.000000  0.974005  157500.000000  1.000000   \n",
      "879   1.000000  1.000000  0.000000  0.000000  112500.000000  1.000000   \n",
      "744   0.000000  0.000000  0.000000  0.000000   67500.000000  1.000000   \n",
      "...        ...       ...       ...       ...            ...       ...   \n",
      "1393  0.878728  0.382983  0.098393  0.000000  202500.000000  1.000000   \n",
      "1344  1.000000  1.000000  0.853176  0.145114  240348.486963  2.731249   \n",
      "527   0.000000  0.000000  1.000000  0.000000  216000.000000  1.000000   \n",
      "1149  0.995074  0.985095  0.000000  0.397428  135000.000000  2.804601   \n",
      "1289  0.000000  1.000000  0.000000  0.673533  135000.000000  1.000000   \n",
      "\n",
      "            6          7         8    9    10        11   12        13  \\\n",
      "788   3.000000  47.293151  3.032877  0.0  0.0  1.000000  0.0  0.000000   \n",
      "803   4.000000  36.093151  6.786301  0.0  0.0  1.000000  0.0  0.000000   \n",
      "1302  2.188169  42.522843  4.968791  0.0  0.0  0.448247  0.0  0.850049   \n",
      "879   2.000000  27.578082  2.284932  0.0  0.0  1.000000  0.0  0.000000   \n",
      "744   2.000000  44.334247  7.745205  0.0  0.0  1.000000  0.0  0.000000   \n",
      "...        ...        ...       ...  ...  ...       ...  ...       ...   \n",
      "1393  2.000000  38.542970  9.297650  0.0  0.0  1.000000  0.0  0.000000   \n",
      "1344  3.916888  47.169620  1.553132  0.0  0.0  0.472443  0.0  0.005004   \n",
      "527   2.000000  42.654795  4.490411  0.0  0.0  1.000000  0.0  0.000000   \n",
      "1149  2.145840  32.157050  5.721969  0.0  0.0  0.140631  0.0  1.000000   \n",
      "1289  2.394543  40.364997  1.753392  0.0  0.0  0.568093  0.0  0.000000   \n",
      "\n",
      "            14        15   16   17        18  \n",
      "788   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "803   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "1302  1.000000  0.000000  0.0  0.0  0.000000  \n",
      "879   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "744   0.000000  0.000000  0.0  0.0  1.000000  \n",
      "...        ...       ...  ...  ...       ...  \n",
      "1393  0.092164  0.000000  0.0  0.0  0.040015  \n",
      "1344  0.146704  0.555787  0.0  0.0  0.000000  \n",
      "527   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "1149  1.000000  0.000000  0.0  0.0  0.000000  \n",
      "1289  1.000000  0.000000  0.0  0.0  0.000000  \n",
      "\n",
      "[1220 rows x 19 columns]\n",
      "\n",
      "Min-Max Scaled Training Data:\n",
      "            0         1         2         3         4         5         6   \\\n",
      "788   0.000000  0.000000  1.000000  0.333333  0.385714  0.250000  0.500000   \n",
      "803   1.000000  1.000000  1.000000  0.666667  0.457143  0.750000  0.750000   \n",
      "1302  0.662759  0.000000  1.000000  0.324668  0.385714  0.250000  0.297042   \n",
      "879   1.000000  1.000000  0.000000  0.000000  0.242857  0.250000  0.250000   \n",
      "744   0.000000  0.000000  0.000000  0.000000  0.100000  0.250000  0.250000   \n",
      "...        ...       ...       ...       ...       ...       ...       ...   \n",
      "1393  0.878728  0.382983  0.098393  0.000000  0.528571  0.250000  0.250000   \n",
      "1344  1.000000  1.000000  0.853176  0.048371  0.648725  0.682812  0.729222   \n",
      "527   0.000000  0.000000  1.000000  0.000000  0.571429  0.250000  0.250000   \n",
      "1149  0.995074  0.985095  0.000000  0.132476  0.314286  0.701150  0.286460   \n",
      "1289  0.000000  1.000000  0.000000  0.224511  0.314286  0.250000  0.348636   \n",
      "\n",
      "            7         8    9    10        11   12        13        14  \\\n",
      "788   0.592168  0.079926  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "803   0.338869  0.185824  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "1302  0.484283  0.134545  0.0  0.0  0.448247  0.0  0.850049  1.000000   \n",
      "879   0.146292  0.058824  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "744   0.525249  0.212878  0.0  0.0  1.000000  0.0  0.000000  0.000000   \n",
      "...        ...       ...  ...  ...       ...  ...       ...       ...   \n",
      "1393  0.394274  0.256678  0.0  0.0  1.000000  0.0  0.000000  0.092164   \n",
      "1344  0.589374  0.038177  0.0  0.0  0.472443  0.0  0.005004  0.146704   \n",
      "527   0.487267  0.121048  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "1149  0.249850  0.155795  0.0  0.0  0.140631  0.0  1.000000  1.000000   \n",
      "1289  0.435481  0.043827  0.0  0.0  0.568093  0.0  0.000000  1.000000   \n",
      "\n",
      "            15   16   17        18  \n",
      "788   0.000000  0.0  0.0  0.000000  \n",
      "803   0.000000  0.0  0.0  0.000000  \n",
      "1302  0.000000  0.0  0.0  0.000000  \n",
      "879   0.000000  0.0  0.0  0.000000  \n",
      "744   0.000000  0.0  0.0  1.000000  \n",
      "...        ...  ...  ...       ...  \n",
      "1393  0.000000  0.0  0.0  0.040015  \n",
      "1344  0.555787  0.0  0.0  0.000000  \n",
      "527   0.000000  0.0  0.0  0.000000  \n",
      "1149  0.000000  0.0  0.0  0.000000  \n",
      "1289  0.000000  0.0  0.0  0.000000  \n",
      "\n",
      "[1220 rows x 19 columns]\n",
      "\n",
      "Original Test Data:\n",
      "        0\n",
      "788   0.0\n",
      "803   0.0\n",
      "1302  1.0\n",
      "879   0.0\n",
      "744   0.0\n",
      "...   ...\n",
      "1393  1.0\n",
      "1344  1.0\n",
      "527   0.0\n",
      "1149  1.0\n",
      "1289  1.0\n",
      "\n",
      "[1220 rows x 1 columns]\n",
      "\n",
      "Min-Max Scaled Test Data:\n",
      "       0   1   2   3   4   5   6   7   8   9   10  11  12  13  14  15  16  17  \\\n",
      "788   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "803   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1302  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "879   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "744   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "...   ...  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..   \n",
      "1393  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1344  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "527   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1149  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1289  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "\n",
      "      18  \n",
      "788  NaN  \n",
      "803  NaN  \n",
      "1302 NaN  \n",
      "879  NaN  \n",
      "744  NaN  \n",
      "...   ..  \n",
      "1393 NaN  \n",
      "1344 NaN  \n",
      "527  NaN  \n",
      "1149 NaN  \n",
      "1289 NaN  \n",
      "\n",
      "[1220 rows x 19 columns]\n"
     ]
    }
   ],
   "source": [
    "scaler = CustomMinMaxScaler()\n",
    "\n",
    "X_train_scaled_new = scaler.fit_transform(X_train)\n",
    "X_test_scaled_new = scaler.transform(X_test)\n",
    "\n",
    "print(\"Original Training Data:\")\n",
    "print(X_train)\n",
    "print(\"\\nMin-Max Scaled Training Data:\")\n",
    "print(X_train_scaled_new)\n",
    "\n",
    "print(\"\\nOriginal Test Data:\")\n",
    "print(X_test)\n",
    "print(\"\\nMin-Max Scaled Test Data:\")\n",
    "print(X_test_scaled_new)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Training Data:\n",
      "            0         1         2         3              4         5   \\\n",
      "788   0.000000  0.000000  1.000000  1.000000  157500.000000  1.000000   \n",
      "803   1.000000  1.000000  1.000000  2.000000  180000.000000  3.000000   \n",
      "1302  0.662759  0.000000  1.000000  0.974005  157500.000000  1.000000   \n",
      "879   1.000000  1.000000  0.000000  0.000000  112500.000000  1.000000   \n",
      "744   0.000000  0.000000  0.000000  0.000000   67500.000000  1.000000   \n",
      "...        ...       ...       ...       ...            ...       ...   \n",
      "1393  0.878728  0.382983  0.098393  0.000000  202500.000000  1.000000   \n",
      "1344  1.000000  1.000000  0.853176  0.145114  240348.486963  2.731249   \n",
      "527   0.000000  0.000000  1.000000  0.000000  216000.000000  1.000000   \n",
      "1149  0.995074  0.985095  0.000000  0.397428  135000.000000  2.804601   \n",
      "1289  0.000000  1.000000  0.000000  0.673533  135000.000000  1.000000   \n",
      "\n",
      "            6          7         8    9    10        11   12        13  \\\n",
      "788   3.000000  47.293151  3.032877  0.0  0.0  1.000000  0.0  0.000000   \n",
      "803   4.000000  36.093151  6.786301  0.0  0.0  1.000000  0.0  0.000000   \n",
      "1302  2.188169  42.522843  4.968791  0.0  0.0  0.448247  0.0  0.850049   \n",
      "879   2.000000  27.578082  2.284932  0.0  0.0  1.000000  0.0  0.000000   \n",
      "744   2.000000  44.334247  7.745205  0.0  0.0  1.000000  0.0  0.000000   \n",
      "...        ...        ...       ...  ...  ...       ...  ...       ...   \n",
      "1393  2.000000  38.542970  9.297650  0.0  0.0  1.000000  0.0  0.000000   \n",
      "1344  3.916888  47.169620  1.553132  0.0  0.0  0.472443  0.0  0.005004   \n",
      "527   2.000000  42.654795  4.490411  0.0  0.0  1.000000  0.0  0.000000   \n",
      "1149  2.145840  32.157050  5.721969  0.0  0.0  0.140631  0.0  1.000000   \n",
      "1289  2.394543  40.364997  1.753392  0.0  0.0  0.568093  0.0  0.000000   \n",
      "\n",
      "            14        15   16   17        18  \n",
      "788   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "803   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "1302  1.000000  0.000000  0.0  0.0  0.000000  \n",
      "879   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "744   0.000000  0.000000  0.0  0.0  1.000000  \n",
      "...        ...       ...  ...  ...       ...  \n",
      "1393  0.092164  0.000000  0.0  0.0  0.040015  \n",
      "1344  0.146704  0.555787  0.0  0.0  0.000000  \n",
      "527   1.000000  0.000000  0.0  0.0  0.000000  \n",
      "1149  1.000000  0.000000  0.0  0.0  0.000000  \n",
      "1289  1.000000  0.000000  0.0  0.0  0.000000  \n",
      "\n",
      "[1220 rows x 19 columns]\n",
      "\n",
      "Min-Max Scaled Training Data:\n",
      "            0         1         2         3         4         5         6   \\\n",
      "788   0.000000  0.000000  1.000000  0.333333  0.385714  0.250000  0.500000   \n",
      "803   1.000000  1.000000  1.000000  0.666667  0.457143  0.750000  0.750000   \n",
      "1302  0.662759  0.000000  1.000000  0.324668  0.385714  0.250000  0.297042   \n",
      "879   1.000000  1.000000  0.000000  0.000000  0.242857  0.250000  0.250000   \n",
      "744   0.000000  0.000000  0.000000  0.000000  0.100000  0.250000  0.250000   \n",
      "...        ...       ...       ...       ...       ...       ...       ...   \n",
      "1393  0.878728  0.382983  0.098393  0.000000  0.528571  0.250000  0.250000   \n",
      "1344  1.000000  1.000000  0.853176  0.048371  0.648725  0.682812  0.729222   \n",
      "527   0.000000  0.000000  1.000000  0.000000  0.571429  0.250000  0.250000   \n",
      "1149  0.995074  0.985095  0.000000  0.132476  0.314286  0.701150  0.286460   \n",
      "1289  0.000000  1.000000  0.000000  0.224511  0.314286  0.250000  0.348636   \n",
      "\n",
      "            7         8    9    10        11   12        13        14  \\\n",
      "788   0.592168  0.079926  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "803   0.338869  0.185824  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "1302  0.484283  0.134545  0.0  0.0  0.448247  0.0  0.850049  1.000000   \n",
      "879   0.146292  0.058824  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "744   0.525249  0.212878  0.0  0.0  1.000000  0.0  0.000000  0.000000   \n",
      "...        ...       ...  ...  ...       ...  ...       ...       ...   \n",
      "1393  0.394274  0.256678  0.0  0.0  1.000000  0.0  0.000000  0.092164   \n",
      "1344  0.589374  0.038177  0.0  0.0  0.472443  0.0  0.005004  0.146704   \n",
      "527   0.487267  0.121048  0.0  0.0  1.000000  0.0  0.000000  1.000000   \n",
      "1149  0.249850  0.155795  0.0  0.0  0.140631  0.0  1.000000  1.000000   \n",
      "1289  0.435481  0.043827  0.0  0.0  0.568093  0.0  0.000000  1.000000   \n",
      "\n",
      "            15   16   17        18  \n",
      "788   0.000000  0.0  0.0  0.000000  \n",
      "803   0.000000  0.0  0.0  0.000000  \n",
      "1302  0.000000  0.0  0.0  0.000000  \n",
      "879   0.000000  0.0  0.0  0.000000  \n",
      "744   0.000000  0.0  0.0  1.000000  \n",
      "...        ...  ...  ...       ...  \n",
      "1393  0.000000  0.0  0.0  0.040015  \n",
      "1344  0.555787  0.0  0.0  0.000000  \n",
      "527   0.000000  0.0  0.0  0.000000  \n",
      "1149  0.000000  0.0  0.0  0.000000  \n",
      "1289  0.000000  0.0  0.0  0.000000  \n",
      "\n",
      "[1220 rows x 19 columns]\n",
      "\n",
      "Original Test Data:\n",
      "        0\n",
      "788   0.0\n",
      "803   0.0\n",
      "1302  1.0\n",
      "879   0.0\n",
      "744   0.0\n",
      "...   ...\n",
      "1393  1.0\n",
      "1344  1.0\n",
      "527   0.0\n",
      "1149  1.0\n",
      "1289  1.0\n",
      "\n",
      "[1220 rows x 1 columns]\n",
      "\n",
      "Min-Max Scaled Test Data:\n",
      "       0   1   2   3   4   5   6   7   8   9   10  11  12  13  14  15  16  17  \\\n",
      "788   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "803   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1302  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "879   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "744   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "...   ...  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..  ..   \n",
      "1393  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1344  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "527   0.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1149  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "1289  1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN   \n",
      "\n",
      "      18  \n",
      "788  NaN  \n",
      "803  NaN  \n",
      "1302 NaN  \n",
      "879  NaN  \n",
      "744  NaN  \n",
      "...   ..  \n",
      "1393 NaN  \n",
      "1344 NaN  \n",
      "527  NaN  \n",
      "1149 NaN  \n",
      "1289 NaN  \n",
      "\n",
      "[1220 rows x 19 columns]\n"
     ]
    }
   ],
   "source": [
    "scaler = CustomMinMaxScaler()\n",
    "\n",
    "X_train_scaled_newX = scaler.fit_transform(X_trainX)\n",
    "X_test_scaled_newX = scaler.transform(X_testX)\n",
    "\n",
    "print(\"Original Training Data:\")\n",
    "print(X_train)\n",
    "print(\"\\nMin-Max Scaled Training Data:\")\n",
    "print(X_train_scaled_new)\n",
    "\n",
    "print(\"\\nOriginal Test Data:\")\n",
    "print(X_test)\n",
    "print(\"\\nMin-Max Scaled Test Data:\")\n",
    "print(X_test_scaled_new)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train_scaled_newX.shape: (848, 19)\n",
      "Y_train.shape: (848,)\n",
      "X_test_scaled_newX.shape: (677, 19)\n",
      "Y_test.shape: (212,)\n"
     ]
    },
    {
     "ename": "AssertionError",
     "evalue": "Inconsistent shapes for test set",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 54\u001b[0m line \u001b[0;36m8\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y151sZmlsZQ%3D%3D?line=5'>6</a>\u001b[0m \u001b[39m# Ensure the shapes are consistent\u001b[39;00m\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y151sZmlsZQ%3D%3D?line=6'>7</a>\u001b[0m \u001b[39massert\u001b[39;00m X_train_scaled_newX\u001b[39m.\u001b[39mshape[\u001b[39m0\u001b[39m] \u001b[39m==\u001b[39m \u001b[39mlen\u001b[39m(Y_train), \u001b[39m\"\u001b[39m\u001b[39mInconsistent shapes for training set\u001b[39m\u001b[39m\"\u001b[39m\n\u001b[0;32m----> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y151sZmlsZQ%3D%3D?line=7'>8</a>\u001b[0m \u001b[39massert\u001b[39;00m X_test_scaled_newX\u001b[39m.\u001b[39mshape[\u001b[39m0\u001b[39m] \u001b[39m==\u001b[39m \u001b[39mlen\u001b[39m(Y_test), \u001b[39m\"\u001b[39m\u001b[39mInconsistent shapes for test set\u001b[39m\u001b[39m\"\u001b[39m\n",
      "\u001b[0;31mAssertionError\u001b[0m: Inconsistent shapes for test set"
     ]
    }
   ],
   "source": [
    "print(\"X_train_scaled_newX.shape:\", X_train_scaled_newX.shape)\n",
    "print(\"Y_train.shape:\", Y_train.shape)\n",
    "print(\"X_test_scaled_newX.shape:\", X_test_scaled_newX.shape)\n",
    "print(\"Y_test.shape:\", Y_test.shape)\n",
    "\n",
    "# Ensure the shapes are consistent\n",
    "assert X_train_scaled_newX.shape[0] == len(Y_train), \"Inconsistent shapes for training set\"\n",
    "assert X_test_scaled_newX.shape[0] == len(Y_test), \"Inconsistent shapes for test set\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score ,roc_auc_score\n",
    "\n",
    "# Initialize the classifiers\n",
    "log_reg = LogisticRegression(max_iter=1000, random_state=10)\n",
    "dtc = DecisionTreeClassifier(random_state=10)\n",
    "rfc = RandomForestClassifier(random_state=42)\n",
    "xgb = XGBClassifier(random_state=10, use_label_encoder=False, eval_metric='logloss')\n",
    "svc = SVC(random_state=10)\n",
    "knn = KNeighborsClassifier()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Found input variables with inconsistent numbers of samples: [212, 677]",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 56\u001b[0m line \u001b[0;36m1\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y150sZmlsZQ%3D%3D?line=13'>14</a>\u001b[0m clf\u001b[39m.\u001b[39mfit(X_train_scaled_newX, Y_train)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y150sZmlsZQ%3D%3D?line=14'>15</a>\u001b[0m y_pred \u001b[39m=\u001b[39m clf\u001b[39m.\u001b[39mpredict(X_test_scaled_newX)\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y150sZmlsZQ%3D%3D?line=16'>17</a>\u001b[0m accuracy \u001b[39m=\u001b[39m accuracy_score(Y_test, y_pred)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y150sZmlsZQ%3D%3D?line=17'>18</a>\u001b[0m precision \u001b[39m=\u001b[39m precision_score(Y_test, y_pred)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y150sZmlsZQ%3D%3D?line=18'>19</a>\u001b[0m recall \u001b[39m=\u001b[39m recall_score(Y_test, y_pred)\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/utils/_param_validation.py:211\u001b[0m, in \u001b[0;36mvalidate_params.<locals>.decorator.<locals>.wrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    205\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[1;32m    206\u001b[0m     \u001b[39mwith\u001b[39;00m config_context(\n\u001b[1;32m    207\u001b[0m         skip_parameter_validation\u001b[39m=\u001b[39m(\n\u001b[1;32m    208\u001b[0m             prefer_skip_nested_validation \u001b[39mor\u001b[39;00m global_skip_validation\n\u001b[1;32m    209\u001b[0m         )\n\u001b[1;32m    210\u001b[0m     ):\n\u001b[0;32m--> 211\u001b[0m         \u001b[39mreturn\u001b[39;00m func(\u001b[39m*\u001b[39margs, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwargs)\n\u001b[1;32m    212\u001b[0m \u001b[39mexcept\u001b[39;00m InvalidParameterError \u001b[39mas\u001b[39;00m e:\n\u001b[1;32m    213\u001b[0m     \u001b[39m# When the function is just a wrapper around an estimator, we allow\u001b[39;00m\n\u001b[1;32m    214\u001b[0m     \u001b[39m# the function to delegate validation to the estimator, but we replace\u001b[39;00m\n\u001b[1;32m    215\u001b[0m     \u001b[39m# the name of the estimator by the name of the function in the error\u001b[39;00m\n\u001b[1;32m    216\u001b[0m     \u001b[39m# message to avoid confusion.\u001b[39;00m\n\u001b[1;32m    217\u001b[0m     msg \u001b[39m=\u001b[39m re\u001b[39m.\u001b[39msub(\n\u001b[1;32m    218\u001b[0m         \u001b[39mr\u001b[39m\u001b[39m\"\u001b[39m\u001b[39mparameter of \u001b[39m\u001b[39m\\\u001b[39m\u001b[39mw+ must be\u001b[39m\u001b[39m\"\u001b[39m,\n\u001b[1;32m    219\u001b[0m         \u001b[39mf\u001b[39m\u001b[39m\"\u001b[39m\u001b[39mparameter of \u001b[39m\u001b[39m{\u001b[39;00mfunc\u001b[39m.\u001b[39m\u001b[39m__qualname__\u001b[39m\u001b[39m}\u001b[39;00m\u001b[39m must be\u001b[39m\u001b[39m\"\u001b[39m,\n\u001b[1;32m    220\u001b[0m         \u001b[39mstr\u001b[39m(e),\n\u001b[1;32m    221\u001b[0m     )\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/metrics/_classification.py:220\u001b[0m, in \u001b[0;36maccuracy_score\u001b[0;34m(y_true, y_pred, normalize, sample_weight)\u001b[0m\n\u001b[1;32m    154\u001b[0m \u001b[39m\u001b[39m\u001b[39m\"\"\"Accuracy classification score.\u001b[39;00m\n\u001b[1;32m    155\u001b[0m \n\u001b[1;32m    156\u001b[0m \u001b[39mIn multilabel classification, this function computes subset accuracy:\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    216\u001b[0m \u001b[39m0.5\u001b[39;00m\n\u001b[1;32m    217\u001b[0m \u001b[39m\"\"\"\u001b[39;00m\n\u001b[1;32m    219\u001b[0m \u001b[39m# Compute accuracy for each possible representation\u001b[39;00m\n\u001b[0;32m--> 220\u001b[0m y_type, y_true, y_pred \u001b[39m=\u001b[39m _check_targets(y_true, y_pred)\n\u001b[1;32m    221\u001b[0m check_consistent_length(y_true, y_pred, sample_weight)\n\u001b[1;32m    222\u001b[0m \u001b[39mif\u001b[39;00m y_type\u001b[39m.\u001b[39mstartswith(\u001b[39m\"\u001b[39m\u001b[39mmultilabel\u001b[39m\u001b[39m\"\u001b[39m):\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/metrics/_classification.py:84\u001b[0m, in \u001b[0;36m_check_targets\u001b[0;34m(y_true, y_pred)\u001b[0m\n\u001b[1;32m     57\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39m_check_targets\u001b[39m(y_true, y_pred):\n\u001b[1;32m     58\u001b[0m \u001b[39m    \u001b[39m\u001b[39m\"\"\"Check that y_true and y_pred belong to the same classification task.\u001b[39;00m\n\u001b[1;32m     59\u001b[0m \n\u001b[1;32m     60\u001b[0m \u001b[39m    This converts multiclass or binary types to a common shape, and raises a\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     82\u001b[0m \u001b[39m    y_pred : array or indicator matrix\u001b[39;00m\n\u001b[1;32m     83\u001b[0m \u001b[39m    \"\"\"\u001b[39;00m\n\u001b[0;32m---> 84\u001b[0m     check_consistent_length(y_true, y_pred)\n\u001b[1;32m     85\u001b[0m     type_true \u001b[39m=\u001b[39m type_of_target(y_true, input_name\u001b[39m=\u001b[39m\u001b[39m\"\u001b[39m\u001b[39my_true\u001b[39m\u001b[39m\"\u001b[39m)\n\u001b[1;32m     86\u001b[0m     type_pred \u001b[39m=\u001b[39m type_of_target(y_pred, input_name\u001b[39m=\u001b[39m\u001b[39m\"\u001b[39m\u001b[39my_pred\u001b[39m\u001b[39m\"\u001b[39m)\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/utils/validation.py:409\u001b[0m, in \u001b[0;36mcheck_consistent_length\u001b[0;34m(*arrays)\u001b[0m\n\u001b[1;32m    407\u001b[0m uniques \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39munique(lengths)\n\u001b[1;32m    408\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mlen\u001b[39m(uniques) \u001b[39m>\u001b[39m \u001b[39m1\u001b[39m:\n\u001b[0;32m--> 409\u001b[0m     \u001b[39mraise\u001b[39;00m \u001b[39mValueError\u001b[39;00m(\n\u001b[1;32m    410\u001b[0m         \u001b[39m\"\u001b[39m\u001b[39mFound input variables with inconsistent numbers of samples: \u001b[39m\u001b[39m%r\u001b[39;00m\u001b[39m\"\u001b[39m\n\u001b[1;32m    411\u001b[0m         \u001b[39m%\u001b[39m [\u001b[39mint\u001b[39m(l) \u001b[39mfor\u001b[39;00m l \u001b[39min\u001b[39;00m lengths]\n\u001b[1;32m    412\u001b[0m     )\n",
      "\u001b[0;31mValueError\u001b[0m: Found input variables with inconsistent numbers of samples: [212, 677]"
     ]
    }
   ],
   "source": [
    "# Dictionary to store results\n",
    "results = {}\n",
    "\n",
    "# List of classifiers\n",
    "classifiers = [('Logistic Regression', log_reg),\n",
    "               ('Decision Tree', dtc),\n",
    "               ('Random Forest', rfc),\n",
    "               ('XGBoost', xgb),\n",
    "               ('Support Vector Machine', svc),\n",
    "               ('K-Nearest Neighbors', knn)]\n",
    "\n",
    "# Train, predict, and store results\n",
    "for name, clf in classifiers:\n",
    "    clf.fit(X_train_scaled_newX, Y_train)\n",
    "    y_pred = clf.predict(X_test_scaled_newX)\n",
    "\n",
    "    accuracy = accuracy_score(Y_test, y_pred)\n",
    "    precision = precision_score(Y_test, y_pred)\n",
    "    recall = recall_score(Y_test, y_pred)\n",
    "    f1 = f1_score(Y_test, y_pred)\n",
    "    roc_auc_test = roc_auc_score(Y_test, y_pred)\n",
    "\n",
    "    results[name] = [accuracy, precision, recall, f1 , roc_auc_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Found input variables with inconsistent numbers of samples: [212, 677]",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 56\u001b[0m line \u001b[0;36m1\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X63sZmlsZQ%3D%3D?line=13'>14</a>\u001b[0m clf\u001b[39m.\u001b[39mfit(X_train_scaled_newX, Y_train)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X63sZmlsZQ%3D%3D?line=14'>15</a>\u001b[0m y_pred \u001b[39m=\u001b[39m clf\u001b[39m.\u001b[39mpredict(X_test_scaled_newX)\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X63sZmlsZQ%3D%3D?line=16'>17</a>\u001b[0m accuracy \u001b[39m=\u001b[39m accuracy_score(Y_test, y_pred)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X63sZmlsZQ%3D%3D?line=17'>18</a>\u001b[0m precision \u001b[39m=\u001b[39m precision_score(Y_test, y_pred)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#X63sZmlsZQ%3D%3D?line=18'>19</a>\u001b[0m recall \u001b[39m=\u001b[39m recall_score(Y_test, y_pred)\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/utils/_param_validation.py:211\u001b[0m, in \u001b[0;36mvalidate_params.<locals>.decorator.<locals>.wrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    205\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[1;32m    206\u001b[0m     \u001b[39mwith\u001b[39;00m config_context(\n\u001b[1;32m    207\u001b[0m         skip_parameter_validation\u001b[39m=\u001b[39m(\n\u001b[1;32m    208\u001b[0m             prefer_skip_nested_validation \u001b[39mor\u001b[39;00m global_skip_validation\n\u001b[1;32m    209\u001b[0m         )\n\u001b[1;32m    210\u001b[0m     ):\n\u001b[0;32m--> 211\u001b[0m         \u001b[39mreturn\u001b[39;00m func(\u001b[39m*\u001b[39margs, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwargs)\n\u001b[1;32m    212\u001b[0m \u001b[39mexcept\u001b[39;00m InvalidParameterError \u001b[39mas\u001b[39;00m e:\n\u001b[1;32m    213\u001b[0m     \u001b[39m# When the function is just a wrapper around an estimator, we allow\u001b[39;00m\n\u001b[1;32m    214\u001b[0m     \u001b[39m# the function to delegate validation to the estimator, but we replace\u001b[39;00m\n\u001b[1;32m    215\u001b[0m     \u001b[39m# the name of the estimator by the name of the function in the error\u001b[39;00m\n\u001b[1;32m    216\u001b[0m     \u001b[39m# message to avoid confusion.\u001b[39;00m\n\u001b[1;32m    217\u001b[0m     msg \u001b[39m=\u001b[39m re\u001b[39m.\u001b[39msub(\n\u001b[1;32m    218\u001b[0m         \u001b[39mr\u001b[39m\u001b[39m\"\u001b[39m\u001b[39mparameter of \u001b[39m\u001b[39m\\\u001b[39m\u001b[39mw+ must be\u001b[39m\u001b[39m\"\u001b[39m,\n\u001b[1;32m    219\u001b[0m         \u001b[39mf\u001b[39m\u001b[39m\"\u001b[39m\u001b[39mparameter of \u001b[39m\u001b[39m{\u001b[39;00mfunc\u001b[39m.\u001b[39m\u001b[39m__qualname__\u001b[39m\u001b[39m}\u001b[39;00m\u001b[39m must be\u001b[39m\u001b[39m\"\u001b[39m,\n\u001b[1;32m    220\u001b[0m         \u001b[39mstr\u001b[39m(e),\n\u001b[1;32m    221\u001b[0m     )\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/metrics/_classification.py:220\u001b[0m, in \u001b[0;36maccuracy_score\u001b[0;34m(y_true, y_pred, normalize, sample_weight)\u001b[0m\n\u001b[1;32m    154\u001b[0m \u001b[39m\u001b[39m\u001b[39m\"\"\"Accuracy classification score.\u001b[39;00m\n\u001b[1;32m    155\u001b[0m \n\u001b[1;32m    156\u001b[0m \u001b[39mIn multilabel classification, this function computes subset accuracy:\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    216\u001b[0m \u001b[39m0.5\u001b[39;00m\n\u001b[1;32m    217\u001b[0m \u001b[39m\"\"\"\u001b[39;00m\n\u001b[1;32m    219\u001b[0m \u001b[39m# Compute accuracy for each possible representation\u001b[39;00m\n\u001b[0;32m--> 220\u001b[0m y_type, y_true, y_pred \u001b[39m=\u001b[39m _check_targets(y_true, y_pred)\n\u001b[1;32m    221\u001b[0m check_consistent_length(y_true, y_pred, sample_weight)\n\u001b[1;32m    222\u001b[0m \u001b[39mif\u001b[39;00m y_type\u001b[39m.\u001b[39mstartswith(\u001b[39m\"\u001b[39m\u001b[39mmultilabel\u001b[39m\u001b[39m\"\u001b[39m):\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/metrics/_classification.py:84\u001b[0m, in \u001b[0;36m_check_targets\u001b[0;34m(y_true, y_pred)\u001b[0m\n\u001b[1;32m     57\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39m_check_targets\u001b[39m(y_true, y_pred):\n\u001b[1;32m     58\u001b[0m \u001b[39m    \u001b[39m\u001b[39m\"\"\"Check that y_true and y_pred belong to the same classification task.\u001b[39;00m\n\u001b[1;32m     59\u001b[0m \n\u001b[1;32m     60\u001b[0m \u001b[39m    This converts multiclass or binary types to a common shape, and raises a\u001b[39;00m\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     82\u001b[0m \u001b[39m    y_pred : array or indicator matrix\u001b[39;00m\n\u001b[1;32m     83\u001b[0m \u001b[39m    \"\"\"\u001b[39;00m\n\u001b[0;32m---> 84\u001b[0m     check_consistent_length(y_true, y_pred)\n\u001b[1;32m     85\u001b[0m     type_true \u001b[39m=\u001b[39m type_of_target(y_true, input_name\u001b[39m=\u001b[39m\u001b[39m\"\u001b[39m\u001b[39my_true\u001b[39m\u001b[39m\"\u001b[39m)\n\u001b[1;32m     86\u001b[0m     type_pred \u001b[39m=\u001b[39m type_of_target(y_pred, input_name\u001b[39m=\u001b[39m\u001b[39m\"\u001b[39m\u001b[39my_pred\u001b[39m\u001b[39m\"\u001b[39m)\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/sklearn/utils/validation.py:409\u001b[0m, in \u001b[0;36mcheck_consistent_length\u001b[0;34m(*arrays)\u001b[0m\n\u001b[1;32m    407\u001b[0m uniques \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39munique(lengths)\n\u001b[1;32m    408\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mlen\u001b[39m(uniques) \u001b[39m>\u001b[39m \u001b[39m1\u001b[39m:\n\u001b[0;32m--> 409\u001b[0m     \u001b[39mraise\u001b[39;00m \u001b[39mValueError\u001b[39;00m(\n\u001b[1;32m    410\u001b[0m         \u001b[39m\"\u001b[39m\u001b[39mFound input variables with inconsistent numbers of samples: \u001b[39m\u001b[39m%r\u001b[39;00m\u001b[39m\"\u001b[39m\n\u001b[1;32m    411\u001b[0m         \u001b[39m%\u001b[39m [\u001b[39mint\u001b[39m(l) \u001b[39mfor\u001b[39;00m l \u001b[39min\u001b[39;00m lengths]\n\u001b[1;32m    412\u001b[0m     )\n",
      "\u001b[0;31mValueError\u001b[0m: Found input variables with inconsistent numbers of samples: [212, 677]"
     ]
    }
   ],
   "source": [
    "# Dictionary to store results\n",
    "results = {}\n",
    "\n",
    "# List of classifiers\n",
    "classifiers = [('Logistic Regression', log_reg),\n",
    "               ('Decision Tree', dtc),\n",
    "               ('Random Forest', rfc),\n",
    "               ('XGBoost', xgb),\n",
    "               ('Support Vector Machine', svc),\n",
    "               ('K-Nearest Neighbors', knn)]\n",
    "\n",
    "# Train, predict, and store results\n",
    "for name, clf in classifiers:\n",
    "    clf.fit(X_train_scaled_newX, Y_train)\n",
    "    y_pred = clf.predict(X_test_scaled_newX)\n",
    "\n",
    "    accuracy = accuracy_score(Y_test, y_pred)\n",
    "    precision = precision_score(Y_test, y_pred)\n",
    "    recall = recall_score(Y_test, y_pred)\n",
    "    f1 = f1_score(Y_test, y_pred)\n",
    "    roc_auc_test = roc_auc_score(Y_test, y_pred)\n",
    "\n",
    "    results[name] = [accuracy, precision, recall, f1 , roc_auc_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "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>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "      <th>F1-Score</th>\n",
       "      <th>ROC AUC Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Logistic Regression</th>\n",
       "      <td>0.910377</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Decision Tree</th>\n",
       "      <td>0.863208</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.263158</td>\n",
       "      <td>0.256410</td>\n",
       "      <td>0.592719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest</th>\n",
       "      <td>0.915094</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.052632</td>\n",
       "      <td>0.100000</td>\n",
       "      <td>0.526316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XGBoost</th>\n",
       "      <td>0.915094</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>0.210526</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.597491</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Support Vector Machine</th>\n",
       "      <td>0.910377</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>K-Nearest Neighbors</th>\n",
       "      <td>0.905660</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.497409</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        Accuracy  Precision    Recall  F1-Score  ROC AUC Score\n",
       "Logistic Regression     0.910377   0.000000  0.000000  0.000000       0.500000\n",
       "Decision Tree           0.863208   0.250000  0.263158  0.256410       0.592719\n",
       "Random Forest           0.915094   1.000000  0.052632  0.100000       0.526316\n",
       "XGBoost                 0.915094   0.571429  0.210526  0.307692       0.597491\n",
       "Support Vector Machine  0.910377   0.000000  0.000000  0.000000       0.500000\n",
       "K-Nearest Neighbors     0.905660   0.000000  0.000000  0.000000       0.497409"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_df = pd.DataFrame(results, index=['Accuracy', 'Precision', 'Recall', 'F1-Score', 'ROC AUC Score']).T\n",
    "results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [],
   "source": [
    "class LogisticRegression:\n",
    "    def __init__(self, learning_rate=0.01, num_iterations=1000, regularization_param=0.1):\n",
    "        self.learning_rate = learning_rate\n",
    "        self.num_iterations = num_iterations\n",
    "        self.regularization_param = regularization_param\n",
    "        self.weights = None\n",
    "        self.bias = None\n",
    "\n",
    "    def _sigmoid(self, z):\n",
    "        return 1 / (1 + np.exp(-z))\n",
    "\n",
    "    def _compute_cost(self, X, y):\n",
    "        m = len(y)\n",
    "        h = self._sigmoid(np.dot(X, self.weights) + self.bias)\n",
    "        cost = (-1 / m) * np.sum(y * np.log(h) + (1 - y) * np.log(1 - h))\n",
    "        reg_term = (self.regularization_param / (2 * m)) * np.sum(self.weights**2)\n",
    "        return cost + reg_term\n",
    "\n",
    "    def _compute_gradients(self, X, y):\n",
    "        m = len(y)\n",
    "        h = self._sigmoid(np.dot(X, self.weights) + self.bias)\n",
    "        error = h - y\n",
    "        gradients = {\n",
    "            'weights': (1 / m) * np.dot(X.T, error) + (self.regularization_param / m) * self.weights,\n",
    "            'bias': np.sum(error) / m\n",
    "        }\n",
    "        return gradients\n",
    "\n",
    "    def fit(self, X, y):\n",
    "        m, n = X.shape\n",
    "        self.weights = np.zeros(n)\n",
    "        self.bias = 0\n",
    "\n",
    "        for _ in range(self.num_iterations):\n",
    "            cost = self._compute_cost(X, y)\n",
    "            gradients = self._compute_gradients(X, y)\n",
    "\n",
    "            self.weights -= self.learning_rate * gradients['weights']\n",
    "            self.bias -= self.learning_rate * gradients['bias']\n",
    "\n",
    "    def predict(self, X):\n",
    "        predictions = self._sigmoid(np.dot(X, self.weights) + self.bias)\n",
    "        return (predictions >= 0.5).astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "operands could not be broadcast together with shapes (848,) (1220,) ",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 53\u001b[0m line \u001b[0;36m2\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=0'>1</a>\u001b[0m model \u001b[39m=\u001b[39m LogisticRegression()\n\u001b[0;32m----> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=1'>2</a>\u001b[0m model\u001b[39m.\u001b[39mfit(X_train, Y_train)\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=3'>4</a>\u001b[0m \u001b[39m# Assuming X_test is your test data\u001b[39;00m\n\u001b[1;32m      <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=4'>5</a>\u001b[0m predictions \u001b[39m=\u001b[39m model\u001b[39m.\u001b[39mpredict(X_test)\n",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 53\u001b[0m line \u001b[0;36m3\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=31'>32</a>\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mbias \u001b[39m=\u001b[39m \u001b[39m0\u001b[39m\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=33'>34</a>\u001b[0m \u001b[39mfor\u001b[39;00m _ \u001b[39min\u001b[39;00m \u001b[39mrange\u001b[39m(\u001b[39mself\u001b[39m\u001b[39m.\u001b[39mnum_iterations):\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=34'>35</a>\u001b[0m     cost \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_compute_cost(X, y)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=35'>36</a>\u001b[0m     gradients \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_compute_gradients(X, y)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=37'>38</a>\u001b[0m     \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mweights \u001b[39m-\u001b[39m\u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mlearning_rate \u001b[39m*\u001b[39m gradients[\u001b[39m'\u001b[39m\u001b[39mweights\u001b[39m\u001b[39m'\u001b[39m]\n",
      "\u001b[1;32m/Users/biryani/Documents/Machine Learning with Python/ml-proj/credit-draft.ipynb Cell 53\u001b[0m line \u001b[0;36m1\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=12'>13</a>\u001b[0m m \u001b[39m=\u001b[39m \u001b[39mlen\u001b[39m(y)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=13'>14</a>\u001b[0m h \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_sigmoid(np\u001b[39m.\u001b[39mdot(X, \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mweights) \u001b[39m+\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mbias)\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=14'>15</a>\u001b[0m cost \u001b[39m=\u001b[39m (\u001b[39m-\u001b[39m\u001b[39m1\u001b[39m \u001b[39m/\u001b[39m m) \u001b[39m*\u001b[39m np\u001b[39m.\u001b[39msum(y \u001b[39m*\u001b[39m np\u001b[39m.\u001b[39mlog(h) \u001b[39m+\u001b[39m (\u001b[39m1\u001b[39m \u001b[39m-\u001b[39m y) \u001b[39m*\u001b[39m np\u001b[39m.\u001b[39mlog(\u001b[39m1\u001b[39m \u001b[39m-\u001b[39m h))\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=15'>16</a>\u001b[0m reg_term \u001b[39m=\u001b[39m (\u001b[39mself\u001b[39m\u001b[39m.\u001b[39mregularization_param \u001b[39m/\u001b[39m (\u001b[39m2\u001b[39m \u001b[39m*\u001b[39m m)) \u001b[39m*\u001b[39m np\u001b[39m.\u001b[39msum(\u001b[39mself\u001b[39m\u001b[39m.\u001b[39mweights\u001b[39m*\u001b[39m\u001b[39m*\u001b[39m\u001b[39m2\u001b[39m)\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/biryani/Documents/Machine%20Learning%20with%20Python/ml-proj/credit-draft.ipynb#Y126sZmlsZQ%3D%3D?line=16'>17</a>\u001b[0m \u001b[39mreturn\u001b[39;00m cost \u001b[39m+\u001b[39m reg_term\n",
      "\u001b[0;31mValueError\u001b[0m: operands could not be broadcast together with shapes (848,) (1220,) "
     ]
    }
   ],
   "source": [
    "model = LogisticRegression()\n",
    "model.fit(X_train, Y_train)\n",
    "\n",
    "# Assuming X_test is your test data\n",
    "predictions = model.predict(X_test)\n",
    "predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions: [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "X_train_with_bias = np.c_[np.ones((X_train.shape[0], 1)), X_train]\n",
    "X_test_with_bias = np.c_[np.ones((X_test.shape[0], 1)), X_test]\n",
    "\n",
    "# # Create and train the logistic regression model\n",
    "logreg = LogisticRegression(learning_rate=0.01, num_iterations=1000)\n",
    "logreg.fit(X_train_with_bias, Y_train)\n",
    "\n",
    "# Make predictions on the test set\n",
    "predictions = logreg.predict(X_test_with_bias)\n",
    "\n",
    "# Print or use predictions as needed\n",
    "print(\"Predictions:\", predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [],
   "source": [
    "class KNNClassifier:\n",
    "    def __init__(self, k=3):\n",
    "        self.k = k\n",
    "\n",
    "    def fit(self, X_train, y_train):\n",
    "        self.X_train = X_train\n",
    "        self.y_train = y_train\n",
    "\n",
    "    def predict(self, X_test):\n",
    "        predictions = []\n",
    "\n",
    "        for sample in X_test:\n",
    "            distances = np.linalg.norm(self.X_train - sample, axis=1)\n",
    "            nearest_neighbors_indices = np.argsort(distances)[:self.k]\n",
    "            nearest_labels = self.y_train[nearest_neighbors_indices]\n",
    "            predicted_label = np.bincount(nearest_labels).argmax()\n",
    "            predictions.append(predicted_label)\n",
    "\n",
    "        return np.array(predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn = KNNClassifier(k=6)\n",
    "knn.fit(X_train, Y_train)\n",
    "predictions = knn.predict(X_test)\n",
    "predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 91.04%\n",
      "Precision: 0.00%\n"
     ]
    }
   ],
   "source": [
    "accuracy = np.mean(predictions == Y_test)\n",
    "print(f\"Accuracy: {accuracy:.2%}\")\n",
    "\n",
    "# Calculate precision\n",
    "true_positive = np.sum((predictions == 1) & (Y_test == 1))\n",
    "false_positive = np.sum((predictions == 1) & (Y_test == 0))\n",
    "\n",
    "precision = true_positive / (true_positive + false_positive) if (true_positive + false_positive) > 0 else 0\n",
    "print(f\"Precision: {precision:.2%}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}