restart_pillkyu.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "df = pd.read_csv(r\"C:\\Users\\Administrator\\Desktop\\train.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# baseline model 구축"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to calculate missing values by column# Funct \n",
    "def missing_values_table(df):\n",
    "        # Total missing values\n",
    "        mis_val = df.isnull().sum()\n",
    "        \n",
    "        # Percentage of missing values\n",
    "        mis_val_percent = 100 * df.isnull().sum() / len(df)\n",
    "        \n",
    "        # Make a table with the results\n",
    "        mis_val_table = pd.concat([mis_val, mis_val_percent], axis=1)\n",
    "        \n",
    "        # Rename the columns\n",
    "        mis_val_table_ren_columns = mis_val_table.rename(\n",
    "        columns = {0 : 'Missing Values', 1 : '% of Total Values'})\n",
    "        \n",
    "        # Sort the table by percentage of missing descending\n",
    "        mis_val_table_ren_columns = mis_val_table_ren_columns[\n",
    "            mis_val_table_ren_columns.iloc[:,1] != 0].sort_values(\n",
    "        '% of Total Values', ascending=False).round(1)\n",
    "        \n",
    "        # Print some summary information\n",
    "        print (\"Your selected dataframe has \" + str(df.shape[1]) + \" columns.\\n\"      \n",
    "            \"There are \" + str(mis_val_table_ren_columns.shape[0]) +\n",
    "              \" columns that have missing values.\")\n",
    "        \n",
    "        # Return the dataframe with missing information\n",
    "        return mis_val_table_ren_columns\n",
    "def check_missing_per_row(df):\n",
    "    \"\"\"\n",
    "    행별로 결측치 개수를 확인하는 함수\n",
    "    :param df: 입력 데이터프레임\n",
    "    :return: 행별 결측치 개수와 비율을 포함한 데이터프레임\n",
    "    \"\"\"\n",
    "    # 각 행별 결측치 개수 계산\n",
    "    missing_count = df.isnull().sum(axis=1)\n",
    "\n",
    "    # 각 행별 결측치 비율 계산\n",
    "    missing_percentage = (missing_count / df.shape[1]) * 100\n",
    "\n",
    "\n",
    "    return df_with_missing_info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Your selected dataframe has 122 columns.\n",
      "There are 67 columns that have missing values.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Missing Values</th>\n",
       "      <th>% of Total Values</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MEDI</th>\n",
       "      <td>214865</td>\n",
       "      <td>69.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_AVG</th>\n",
       "      <td>214865</td>\n",
       "      <td>69.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MODE</th>\n",
       "      <td>214865</td>\n",
       "      <td>69.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>213514</td>\n",
       "      <td>69.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MODE</th>\n",
       "      <td>213514</td>\n",
       "      <td>69.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_AVG</th>\n",
       "      <td>213514</td>\n",
       "      <td>69.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FONDKAPREMONT_MODE</th>\n",
       "      <td>210295</td>\n",
       "      <td>68.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MODE</th>\n",
       "      <td>210199</td>\n",
       "      <td>68.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>210199</td>\n",
       "      <td>68.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_AVG</th>\n",
       "      <td>210199</td>\n",
       "      <td>68.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_MODE</th>\n",
       "      <td>208642</td>\n",
       "      <td>67.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_MEDI</th>\n",
       "      <td>208642</td>\n",
       "      <td>67.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_AVG</th>\n",
       "      <td>208642</td>\n",
       "      <td>67.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_MODE</th>\n",
       "      <td>204488</td>\n",
       "      <td>66.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_MEDI</th>\n",
       "      <td>204488</td>\n",
       "      <td>66.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_AVG</th>\n",
       "      <td>204488</td>\n",
       "      <td>66.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OWN_CAR_AGE</th>\n",
       "      <td>202929</td>\n",
       "      <td>66.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_AVG</th>\n",
       "      <td>182590</td>\n",
       "      <td>59.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_MEDI</th>\n",
       "      <td>182590</td>\n",
       "      <td>59.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_MODE</th>\n",
       "      <td>182590</td>\n",
       "      <td>59.4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          Missing Values  % of Total Values\n",
       "COMMONAREA_MEDI                   214865               69.9\n",
       "COMMONAREA_AVG                    214865               69.9\n",
       "COMMONAREA_MODE                   214865               69.9\n",
       "NONLIVINGAPARTMENTS_MEDI          213514               69.4\n",
       "NONLIVINGAPARTMENTS_MODE          213514               69.4\n",
       "NONLIVINGAPARTMENTS_AVG           213514               69.4\n",
       "FONDKAPREMONT_MODE                210295               68.4\n",
       "LIVINGAPARTMENTS_MODE             210199               68.4\n",
       "LIVINGAPARTMENTS_MEDI             210199               68.4\n",
       "LIVINGAPARTMENTS_AVG              210199               68.4\n",
       "FLOORSMIN_MODE                    208642               67.8\n",
       "FLOORSMIN_MEDI                    208642               67.8\n",
       "FLOORSMIN_AVG                     208642               67.8\n",
       "YEARS_BUILD_MODE                  204488               66.5\n",
       "YEARS_BUILD_MEDI                  204488               66.5\n",
       "YEARS_BUILD_AVG                   204488               66.5\n",
       "OWN_CAR_AGE                       202929               66.0\n",
       "LANDAREA_AVG                      182590               59.4\n",
       "LANDAREA_MEDI                     182590               59.4\n",
       "LANDAREA_MODE                     182590               59.4"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Missing values statistics\n",
    "missing_values = missing_values_table(df)\n",
    "missing_values.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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>SK_ID_CURR</th>\n",
       "      <th>TARGET</th>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <th>CODE_GENDER</th>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <th>...</th>\n",
       "      <th>FLAG_DOCUMENT_20</th>\n",
       "      <th>FLAG_DOCUMENT_21</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_HOUR</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_DAY</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_WEEK</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_MON</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_QRT</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_YEAR</th>\n",
       "      <th>Missing Count</th>\n",
       "      <th>Missing Percentage</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>185713</th>\n",
       "      <td>315294</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>6750.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133770</th>\n",
       "      <td>255145</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>225000.0</td>\n",
       "      <td>405000.0</td>\n",
       "      <td>20250.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>197736</th>\n",
       "      <td>329262</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>180000.0</td>\n",
       "      <td>9000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116937</th>\n",
       "      <td>235599</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>225000.0</td>\n",
       "      <td>720000.0</td>\n",
       "      <td>36000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269492</th>\n",
       "      <td>412312</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>450000.0</td>\n",
       "      <td>1350000.0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69707</th>\n",
       "      <td>180861</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>10125.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269786</th>\n",
       "      <td>412671</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1</td>\n",
       "      <td>99000.0</td>\n",
       "      <td>247500.0</td>\n",
       "      <td>12375.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150206</th>\n",
       "      <td>274127</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>450000.0</td>\n",
       "      <td>1237500.0</td>\n",
       "      <td>61875.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>244833</th>\n",
       "      <td>383369</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>2</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>270000.0</td>\n",
       "      <td>13500.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61</td>\n",
       "      <td>50.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>290495</th>\n",
       "      <td>436548</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>2</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>157500.0</td>\n",
       "      <td>7875.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127942</th>\n",
       "      <td>248389</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>1</td>\n",
       "      <td>94500.0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>10125.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258474</th>\n",
       "      <td>399118</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>225000.0</td>\n",
       "      <td>1350000.0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12087</th>\n",
       "      <td>114098</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>315000.0</td>\n",
       "      <td>855000.0</td>\n",
       "      <td>42750.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115346</th>\n",
       "      <td>233761</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>90000.0</td>\n",
       "      <td>180000.0</td>\n",
       "      <td>9000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271341</th>\n",
       "      <td>414557</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>144000.0</td>\n",
       "      <td>450000.0</td>\n",
       "      <td>22500.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53941</th>\n",
       "      <td>162499</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>225000.0</td>\n",
       "      <td>337500.0</td>\n",
       "      <td>16875.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87399</th>\n",
       "      <td>201436</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>450000.0</td>\n",
       "      <td>22500.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68013</th>\n",
       "      <td>178873</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>180000.0</td>\n",
       "      <td>9000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>260125</th>\n",
       "      <td>401026</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>382500.0</td>\n",
       "      <td>19125.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172595</th>\n",
       "      <td>300004</td>\n",
       "      <td>1</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>112500.0</td>\n",
       "      <td>270000.0</td>\n",
       "      <td>13500.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60</td>\n",
       "      <td>49.180328</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20 rows × 124 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        SK_ID_CURR  TARGET NAME_CONTRACT_TYPE CODE_GENDER FLAG_OWN_CAR  \\\n",
       "185713      315294       0    Revolving loans           M            N   \n",
       "133770      255145       0    Revolving loans           F            N   \n",
       "197736      329262       0    Revolving loans           M            N   \n",
       "116937      235599       0    Revolving loans           M            N   \n",
       "269492      412312       0    Revolving loans           F            N   \n",
       "69707       180861       0    Revolving loans           F            N   \n",
       "269786      412671       0    Revolving loans           F            N   \n",
       "150206      274127       0    Revolving loans           M            N   \n",
       "244833      383369       0    Revolving loans           F            N   \n",
       "290495      436548       0    Revolving loans           F            N   \n",
       "127942      248389       0    Revolving loans           F            N   \n",
       "258474      399118       0    Revolving loans           M            N   \n",
       "12087       114098       0    Revolving loans           F            N   \n",
       "115346      233761       0    Revolving loans           F            N   \n",
       "271341      414557       0    Revolving loans           F            N   \n",
       "53941       162499       0    Revolving loans           F            Y   \n",
       "87399       201436       0    Revolving loans           F            N   \n",
       "68013       178873       0    Revolving loans           F            N   \n",
       "260125      401026       0    Revolving loans           F            N   \n",
       "172595      300004       1    Revolving loans           F            N   \n",
       "\n",
       "       FLAG_OWN_REALTY  CNT_CHILDREN  AMT_INCOME_TOTAL  AMT_CREDIT  \\\n",
       "185713               N             0           67500.0    135000.0   \n",
       "133770               Y             0          225000.0    405000.0   \n",
       "197736               N             0          135000.0    180000.0   \n",
       "116937               N             0          225000.0    720000.0   \n",
       "269492               N             0          450000.0   1350000.0   \n",
       "69707                Y             0           67500.0    202500.0   \n",
       "269786               N             1           99000.0    247500.0   \n",
       "150206               N             0          450000.0   1237500.0   \n",
       "244833               Y             2           67500.0    270000.0   \n",
       "290495               Y             2          135000.0    157500.0   \n",
       "127942               Y             1           94500.0    202500.0   \n",
       "258474               Y             0          225000.0   1350000.0   \n",
       "12087                Y             0          315000.0    855000.0   \n",
       "115346               N             0           90000.0    180000.0   \n",
       "271341               Y             0          144000.0    450000.0   \n",
       "53941                Y             0          225000.0    337500.0   \n",
       "87399                Y             0          135000.0    450000.0   \n",
       "68013                Y             0           67500.0    180000.0   \n",
       "260125               Y             0          135000.0    382500.0   \n",
       "172595               Y             0          112500.0    270000.0   \n",
       "\n",
       "        AMT_ANNUITY  ...  FLAG_DOCUMENT_20 FLAG_DOCUMENT_21  \\\n",
       "185713       6750.0  ...                 0                0   \n",
       "133770      20250.0  ...                 0                0   \n",
       "197736       9000.0  ...                 0                0   \n",
       "116937      36000.0  ...                 0                0   \n",
       "269492      67500.0  ...                 0                0   \n",
       "69707       10125.0  ...                 0                0   \n",
       "269786      12375.0  ...                 0                0   \n",
       "150206      61875.0  ...                 0                0   \n",
       "244833      13500.0  ...                 0                0   \n",
       "290495       7875.0  ...                 0                0   \n",
       "127942      10125.0  ...                 0                0   \n",
       "258474      67500.0  ...                 0                0   \n",
       "12087       42750.0  ...                 0                0   \n",
       "115346       9000.0  ...                 0                0   \n",
       "271341      22500.0  ...                 0                0   \n",
       "53941       16875.0  ...                 0                0   \n",
       "87399       22500.0  ...                 0                0   \n",
       "68013        9000.0  ...                 0                0   \n",
       "260125      19125.0  ...                 0                0   \n",
       "172595      13500.0  ...                 0                0   \n",
       "\n",
       "       AMT_REQ_CREDIT_BUREAU_HOUR AMT_REQ_CREDIT_BUREAU_DAY  \\\n",
       "185713                        NaN                       NaN   \n",
       "133770                        NaN                       NaN   \n",
       "197736                        NaN                       NaN   \n",
       "116937                        NaN                       NaN   \n",
       "269492                        NaN                       NaN   \n",
       "69707                         NaN                       NaN   \n",
       "269786                        NaN                       NaN   \n",
       "150206                        NaN                       NaN   \n",
       "244833                        NaN                       NaN   \n",
       "290495                        NaN                       NaN   \n",
       "127942                        NaN                       NaN   \n",
       "258474                        NaN                       NaN   \n",
       "12087                         NaN                       NaN   \n",
       "115346                        NaN                       NaN   \n",
       "271341                        NaN                       NaN   \n",
       "53941                         NaN                       NaN   \n",
       "87399                         NaN                       NaN   \n",
       "68013                         NaN                       NaN   \n",
       "260125                        NaN                       NaN   \n",
       "172595                        NaN                       NaN   \n",
       "\n",
       "       AMT_REQ_CREDIT_BUREAU_WEEK AMT_REQ_CREDIT_BUREAU_MON  \\\n",
       "185713                        NaN                       NaN   \n",
       "133770                        NaN                       NaN   \n",
       "197736                        NaN                       NaN   \n",
       "116937                        NaN                       NaN   \n",
       "269492                        NaN                       NaN   \n",
       "69707                         NaN                       NaN   \n",
       "269786                        NaN                       NaN   \n",
       "150206                        NaN                       NaN   \n",
       "244833                        NaN                       NaN   \n",
       "290495                        NaN                       NaN   \n",
       "127942                        NaN                       NaN   \n",
       "258474                        NaN                       NaN   \n",
       "12087                         NaN                       NaN   \n",
       "115346                        NaN                       NaN   \n",
       "271341                        NaN                       NaN   \n",
       "53941                         NaN                       NaN   \n",
       "87399                         NaN                       NaN   \n",
       "68013                         NaN                       NaN   \n",
       "260125                        NaN                       NaN   \n",
       "172595                        NaN                       NaN   \n",
       "\n",
       "        AMT_REQ_CREDIT_BUREAU_QRT  AMT_REQ_CREDIT_BUREAU_YEAR  Missing Count  \\\n",
       "185713                        NaN                         NaN             61   \n",
       "133770                        NaN                         NaN             61   \n",
       "197736                        NaN                         NaN             61   \n",
       "116937                        NaN                         NaN             61   \n",
       "269492                        NaN                         NaN             61   \n",
       "69707                         NaN                         NaN             61   \n",
       "269786                        NaN                         NaN             61   \n",
       "150206                        NaN                         NaN             61   \n",
       "244833                        NaN                         NaN             61   \n",
       "290495                        NaN                         NaN             60   \n",
       "127942                        NaN                         NaN             60   \n",
       "258474                        NaN                         NaN             60   \n",
       "12087                         NaN                         NaN             60   \n",
       "115346                        NaN                         NaN             60   \n",
       "271341                        NaN                         NaN             60   \n",
       "53941                         NaN                         NaN             60   \n",
       "87399                         NaN                         NaN             60   \n",
       "68013                         NaN                         NaN             60   \n",
       "260125                        NaN                         NaN             60   \n",
       "172595                        NaN                         NaN             60   \n",
       "\n",
       "        Missing Percentage  \n",
       "185713           50.000000  \n",
       "133770           50.000000  \n",
       "197736           50.000000  \n",
       "116937           50.000000  \n",
       "269492           50.000000  \n",
       "69707            50.000000  \n",
       "269786           50.000000  \n",
       "150206           50.000000  \n",
       "244833           50.000000  \n",
       "290495           49.180328  \n",
       "127942           49.180328  \n",
       "258474           49.180328  \n",
       "12087            49.180328  \n",
       "115346           49.180328  \n",
       "271341           49.180328  \n",
       "53941            49.180328  \n",
       "87399            49.180328  \n",
       "68013            49.180328  \n",
       "260125           49.180328  \n",
       "172595           49.180328  \n",
       "\n",
       "[20 rows x 124 columns]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "check_row=check_missing_per_row(df)\n",
    "check_row = check_row.sort_values(by='Missing Percentage', key=lambda x: x, ascending=False)\n",
    "check_row.head(20)\n",
    "check_row[check_row['Missing Percentage']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "FONDKAPREMONT_MODE: 주택 수리 자금 형태 -  원핫\n",
    "HOUSETYPE_MODE: 주택 유형 - 타겟\n",
    "WALLSMATERIAL_MODE: 벽 재질 -  타겟\n",
    "EMERGENCYSTATE_MODE: 비상 상태 여부 - 바이너리"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "# 범주형 변수 목록 추출\n",
    "selected_columns = [\n",
    "    'FONDKAPREMONT_MODE',\n",
    "    'HOUSETYPE_MODE',\n",
    "    'WALLSMATERIAL_MODE',\n",
    "    'EMERGENCYSTATE_MODE'\n",
    "]\n",
    "\n",
    "\n",
    "# target 변수 이름 (예: 'target' 변수로 가정)\n",
    "target_column = 'TARGET'\n",
    "\n",
    "# 각 범주형 변수에 대해 분포 시각화 및 1인 값의 비율 계산\n",
    "for column in selected_columns:\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    sns.countplot(data=df, x=column, hue=target_column, palette=\"Set2\")\n",
    "    plt.title(f\"{column}에 따른 {target_column} 분포\")\n",
    "    plt.xlabel(column)\n",
    "    plt.ylabel(\"Count\")\n",
    "    plt.legend(title=target_column)\n",
    "    plt.xticks(rotation=45)\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "    # TARGET이 1인 비율 계산 및 출력\n",
    "    target_1_ratio = df.groupby(column)[target_column].mean()\n",
    "    print(f\"\\n{column}에 따른 {target_column}이 1인 비율:\")\n",
    "    print(target_1_ratio)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "def encode_columns(df):\n",
    "    \"\"\"\n",
    "    주어진 규칙에 따라 특정 열들을 인코딩하는 함수\n",
    "    :param df: 데이터프레임\n",
    "    :return: 인코딩된 데이터프레임\n",
    "    \"\"\"\n",
    "    # 1. NAME_CONTRACT_TYPE: 대출 유형 (현금 또는 리볼빙)\n",
    "    df['NAME_CONTRACT_TYPE'] = df['NAME_CONTRACT_TYPE'].map({'Cash loans': 0, 'Revolving loans': 1})\n",
    "\n",
    "    # 2. CODE_GENDER: 성별 (XNA 제거, 남성: 0, 여성: 1)\n",
    "    df = df[df['CODE_GENDER'] != 'XNA'].copy()\n",
    "    df['CODE_GENDER'] = df['CODE_GENDER'].map({'M': 0, 'F': 1})\n",
    "\n",
    "    # 3. FLAG_OWN_CAR: 자동차 소유 여부 (Y: 1, N: 0)\n",
    "    df.loc[:, 'FLAG_OWN_CAR'] = df['FLAG_OWN_CAR'].map({'Y': 1, 'N': 0})\n",
    "\n",
    "    # 4. FLAG_OWN_REALTY: 주택 소유 여부 (Y: 1, N: 0)\n",
    "    df.loc[:, 'FLAG_OWN_REALTY'] = df['FLAG_OWN_REALTY'].map({'Y': 1, 'N': 0})\n",
    "\n",
    "    # 5. NAME_TYPE_SUITE: 대출 신청 시 동행인 (Unaccompanied: 0, 나머지: 1)\n",
    "    df.loc[:, 'NAME_TYPE_SUITE'] = df['NAME_TYPE_SUITE'].map({'Unaccompanied': 0}).fillna(1)\n",
    "\n",
    "    # 6. NAME_INCOME_TYPE: 소득 유형\n",
    "    income_map = {\n",
    "        'Working': 0,\n",
    "        'Commercial associate': 1,\n",
    "        'Pensioner': 2,\n",
    "        'State servant': 2\n",
    "    }\n",
    "    df['NAME_INCOME_TYPE'] = df['NAME_INCOME_TYPE'].map(income_map).fillna(3)\n",
    "\n",
    "    # 7. NAME_EDUCATION_TYPE: 학력 수준 (순서형 인코딩 0~4)\n",
    "    education_map = {\n",
    "        'Lower secondary': 0,\n",
    "        'Secondary / secondary special': 1,\n",
    "        'Incomplete higher': 2,\n",
    "        'Higher education': 3,\n",
    "        'Academic degree': 4\n",
    "    }\n",
    "    df['NAME_EDUCATION_TYPE'] = df['NAME_EDUCATION_TYPE'].map(education_map)\n",
    "\n",
    "    # 8. NAME_FAMILY_STATUS: 가족 상태 (Married or Civil marriage → 1, Single → 2, Separated or Widow → 3)\n",
    "    family_status_map = {\n",
    "        'Married': 1, 'Civil marriage': 1,\n",
    "        'Single / not married': 2,\n",
    "        'Separated': 3, 'Widow': 3\n",
    "    }\n",
    "    df['NAME_FAMILY_STATUS'] = df['NAME_FAMILY_STATUS'].map(family_status_map)\n",
    "\n",
    "    # 9. NAME_HOUSING_TYPE: 주거 형태 (House/apartment: 1, 나머지: 0)\n",
    "    df['NAME_HOUSING_TYPE'] = df['NAME_HOUSING_TYPE'].map({\n",
    "        'House / apartment': 1\n",
    "    }).fillna(0)\n",
    "\n",
    "    # 10. OCCUPATION_TYPE: 직업 유형 (Target Encoding)\n",
    "    if 'OCCUPATION_TYPE' in df.columns and 'TARGET' in df.columns:\n",
    "        occupation_target_map = df.groupby('OCCUPATION_TYPE')['TARGET'].mean()\n",
    "        df['OCCUPATION_TYPE'] = df['OCCUPATION_TYPE'].map(occupation_target_map)\n",
    "\n",
    "    # 11. ORGANIZATION_TYPE: 대출 신청자의 소속 조직 유형 (Frequency Encoding)\n",
    "    if 'ORGANIZATION_TYPE' in df.columns:\n",
    "        organization_type_map = df['ORGANIZATION_TYPE'].value_counts() / len(df)\n",
    "        df['ORGANIZATION_TYPE'] = df['ORGANIZATION_TYPE'].map(organization_type_map)\n",
    "\n",
    "    # 12. WEEKDAY_APPR_PROCESS_START: 평일(0), 주말(1)로 인코딩\n",
    "    if 'WEEKDAY_APPR_PROCESS_START' in df.columns:\n",
    "        weekend_days = {'SATURDAY', 'SUNDAY'}\n",
    "        df['WEEKDAY_APPR_PROCESS_START'] = df['WEEKDAY_APPR_PROCESS_START'].apply(\n",
    "            lambda x: 1 if x in weekend_days else 0\n",
    "        )\n",
    "\n",
    "    # 13. FONDKAPREMONT_MODE: 원핫 인코딩\n",
    "    df = pd.get_dummies(df, columns=['FONDKAPREMONT_MODE'], drop_first=True)\n",
    "\n",
    "    # 14. HOUSETYPE_MODE: 타겟 인코딩\n",
    "    if 'HOUSETYPE_MODE' in df.columns and 'TARGET' in df.columns:\n",
    "        house_target_mean = df.groupby('HOUSETYPE_MODE')['TARGET'].mean()\n",
    "        df['HOUSETYPE_MODE_ENCODED'] = df['HOUSETYPE_MODE'].map(house_target_mean)\n",
    "        df.drop(columns=['HOUSETYPE_MODE'], inplace=True)\n",
    "\n",
    "    # 15. WALLSMATERIAL_MODE: 타겟 인코딩\n",
    "    if 'WALLSMATERIAL_MODE' in df.columns and 'TARGET' in df.columns:\n",
    "        walls_target_mean = df.groupby('WALLSMATERIAL_MODE')['TARGET'].mean()\n",
    "        df['WALLSMATERIAL_MODE_ENCODED'] = df['WALLSMATERIAL_MODE'].map(walls_target_mean)\n",
    "        df.drop(columns=['WALLSMATERIAL_MODE'], inplace=True)\n",
    "\n",
    "    # 16. EMERGENCYSTATE_MODE: 바이너리 인코딩 (Yes -> 1, No -> 0)\n",
    "    if 'EMERGENCYSTATE_MODE' in df.columns:\n",
    "        df['EMERGENCYSTATE_MODE'] = df['EMERGENCYSTATE_MODE'].map({'Yes': 1, 'No': 0})\n",
    "\n",
    "    # NAME_INCOME_TYPE 및 NAME_FAMILY_STATUS를 원핫 인코딩\n",
    "    df = pd.get_dummies(df, columns=['NAME_INCOME_TYPE', 'NAME_FAMILY_STATUS'], drop_first=True)\n",
    "    \n",
    "    \n",
    "\n",
    "    return df\n",
    "\n",
    "\n",
    "# 데이터프레임 예시\n",
    "df = encode_columns(df)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 변환할 열 목록\n",
    "columns_to_convert = [\n",
    "    'NAME_INCOME_TYPE_1.0',\n",
    "    'NAME_INCOME_TYPE_2.0',\n",
    "    'NAME_INCOME_TYPE_3.0',\n",
    "    'NAME_FAMILY_STATUS_2.0',\n",
    "    'NAME_FAMILY_STATUS_3.0'\n",
    "]\n",
    "for column in df:\n",
    "        if column in df.columns:\n",
    "            df[column] = df[column].apply(lambda x: int(x) if pd.notnull(x) and isinstance(x, bool) else x)  # True -> 1, False -> 0, 결측치는 유지\n",
    "        else:\n",
    "            print(f\"Warning: Column '{column}' not found in the dataframe.\")\n",
    "df.drop(columns='SK_ID_CURR', inplace=True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8940802"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 결측치를 중앙값으로 대체하였습니다"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "# 예제 데이터프레임\n",
    "# df = pd.read_csv('your_data.csv')\n",
    "\n",
    "# 각 클래스별로 결측치를 중앙값으로 대체\n",
    "def impute_by_class(df, target_col):\n",
    "    imputed_df = pd.DataFrame()\n",
    "    for cls in df[target_col].unique():  # 각 클래스에 대해\n",
    "        class_data = df[df[target_col] == cls]\n",
    "        class_data = class_data.fillna(class_data.median())  # 중앙값으로 결측치 대체\n",
    "        imputed_df = pd.concat([imputed_df, class_data], axis=0)\n",
    "    return imputed_df\n",
    "\n",
    "# 결측치 처리\n",
    "df_imputed = impute_by_class(df, target_col='TARGET')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_imputed.isnull().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "열: TARGET, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: NAME_CONTRACT_TYPE, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: CODE_GENDER, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_OWN_CAR, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_OWN_REALTY, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: CNT_CHILDREN, 왜도: 1.97, 첨도: 7.90, 스케일링 방법: RobustScaler\n",
      "열: AMT_INCOME_TOTAL, 왜도: 391.56, 첨도: 191784.26, 스케일링 방법: RobustScaler\n",
      "열: AMT_CREDIT, 왜도: 1.23, 첨도: 1.93, 스케일링 방법: RobustScaler\n",
      "열: AMT_ANNUITY, 왜도: 1.58, 첨도: 7.71, 스케일링 방법: RobustScaler\n",
      "열: AMT_GOODS_PRICE, 왜도: 1.35, 첨도: 2.44, 스케일링 방법: RobustScaler\n",
      "열: NAME_TYPE_SUITE, 고유값: [0. 1.], 스케일링 제외 (바이너리 데이터)\n",
      "열: NAME_EDUCATION_TYPE, 왜도: 1.02, 첨도: -0.72, 스케일링 방법: RobustScaler\n",
      "열: NAME_HOUSING_TYPE, 고유값: [1. 0.], 스케일링 제외 (바이너리 데이터)\n",
      "열: REGION_POPULATION_RELATIVE, 왜도: 1.49, 첨도: 3.26, 스케일링 방법: RobustScaler\n",
      "열: DAYS_BIRTH, 왜도: -0.12, 첨도: -1.05, 스케일링 방법: StandardScaler\n",
      "열: DAYS_EMPLOYED, 왜도: 1.66, 첨도: 0.77, 스케일링 방법: RobustScaler\n",
      "열: DAYS_REGISTRATION, 왜도: -0.59, 첨도: -0.32, 스케일링 방법: RobustScaler\n",
      "열: DAYS_ID_PUBLISH, 왜도: 0.35, 첨도: -1.11, 스케일링 방법: StandardScaler\n",
      "열: OWN_CAR_AGE, 왜도: 5.20, 첨도: 35.18, 스케일링 방법: RobustScaler\n",
      "열: FLAG_MOBIL, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_EMP_PHONE, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_WORK_PHONE, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_CONT_MOBILE, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_PHONE, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_EMAIL, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: OCCUPATION_TYPE, 왜도: -0.29, 첨도: 0.82, 스케일링 방법: StandardScaler\n",
      "열: CNT_FAM_MEMBERS, 왜도: 0.99, 첨도: 2.80, 스케일링 방법: RobustScaler\n",
      "열: REGION_RATING_CLIENT, 왜도: 0.09, 첨도: 0.80, 스케일링 방법: StandardScaler\n",
      "열: REGION_RATING_CLIENT_W_CITY, 왜도: 0.06, 첨도: 0.93, 스케일링 방법: StandardScaler\n",
      "열: WEEKDAY_APPR_PROCESS_START, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: HOUR_APPR_PROCESS_START, 왜도: -0.03, 첨도: -0.19, 스케일링 방법: StandardScaler\n",
      "열: REG_REGION_NOT_LIVE_REGION, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: REG_REGION_NOT_WORK_REGION, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: LIVE_REGION_NOT_WORK_REGION, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: REG_CITY_NOT_LIVE_CITY, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: REG_CITY_NOT_WORK_CITY, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: LIVE_CITY_NOT_WORK_CITY, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: ORGANIZATION_TYPE, 왜도: 0.11, 첨도: -1.69, 스케일링 방법: StandardScaler\n",
      "열: EXT_SOURCE_1, 왜도: -0.17, 첨도: 1.24, 스케일링 방법: StandardScaler\n",
      "열: EXT_SOURCE_2, 왜도: -0.80, 첨도: -0.26, 스케일링 방법: RobustScaler\n",
      "열: EXT_SOURCE_3, 왜도: -0.52, 첨도: -0.14, 스케일링 방법: RobustScaler\n",
      "열: APARTMENTS_AVG, 왜도: 4.12, 첨도: 27.11, 스케일링 방법: RobustScaler\n",
      "열: BASEMENTAREA_AVG, 왜도: 5.83, 첨도: 67.95, 스케일링 방법: RobustScaler\n",
      "열: YEARS_BEGINEXPLUATATION_AVG, 왜도: -21.74, 첨도: 489.14, 스케일링 방법: RobustScaler\n",
      "열: YEARS_BUILD_AVG, 왜도: -1.73, 첨도: 19.20, 스케일링 방법: RobustScaler\n",
      "열: COMMONAREA_AVG, 왜도: 10.10, 첨도: 154.34, 스케일링 방법: RobustScaler\n",
      "열: ELEVATORS_AVG, 왜도: 3.85, 첨도: 19.20, 스케일링 방법: RobustScaler\n",
      "열: ENTRANCES_AVG, 왜도: 3.62, 첨도: 27.16, 스케일링 방법: RobustScaler\n",
      "열: FLOORSMAX_AVG, 왜도: 2.30, 첨도: 8.33, 스케일링 방법: RobustScaler\n",
      "열: FLOORSMIN_AVG, 왜도: 2.16, 첨도: 11.43, 스케일링 방법: RobustScaler\n",
      "열: LANDAREA_AVG, 왜도: 7.29, 첨도: 90.31, 스케일링 방법: RobustScaler\n",
      "열: LIVINGAPARTMENTS_AVG, 왜도: 5.94, 첨도: 59.58, 스케일링 방법: RobustScaler\n",
      "열: LIVINGAREA_AVG, 왜도: 4.38, 첨도: 28.57, 스케일링 방법: RobustScaler\n",
      "열: NONLIVINGAPARTMENTS_AVG, 왜도: 27.82, 첨도: 920.45, 스케일링 방법: RobustScaler\n",
      "열: NONLIVINGAREA_AVG, 왜도: 9.65, 첨도: 139.77, 스케일링 방법: RobustScaler\n",
      "열: APARTMENTS_MODE, 왜도: 4.21, 첨도: 27.79, 스케일링 방법: RobustScaler\n",
      "열: BASEMENTAREA_MODE, 왜도: 5.71, 첨도: 64.36, 스케일링 방법: RobustScaler\n",
      "열: YEARS_BEGINEXPLUATATION_MODE, 왜도: -20.69, 첨도: 434.14, 스케일링 방법: RobustScaler\n",
      "열: YEARS_BUILD_MODE, 왜도: -1.86, 첨도: 20.39, 스케일링 방법: RobustScaler\n",
      "열: COMMONAREA_MODE, 왜도: 10.36, 첨도: 162.61, 스케일링 방법: RobustScaler\n",
      "열: ELEVATORS_MODE, 왜도: 4.00, 첨도: 20.84, 스케일링 방법: RobustScaler\n",
      "열: ENTRANCES_MODE, 왜도: 3.54, 첨도: 26.61, 스케일링 방법: RobustScaler\n",
      "열: FLOORSMAX_MODE, 왜도: 2.31, 첨도: 8.58, 스케일링 방법: RobustScaler\n",
      "열: FLOORSMIN_MODE, 왜도: 2.11, 첨도: 11.37, 스케일링 방법: RobustScaler\n",
      "열: LANDAREA_MODE, 왜도: 7.17, 첨도: 86.60, 스케일링 방법: RobustScaler\n",
      "열: LIVINGAPARTMENTS_MODE, 왜도: 5.73, 첨도: 52.75, 스케일링 방법: RobustScaler\n",
      "열: LIVINGAREA_MODE, 왜도: 4.45, 첨도: 28.88, 스케일링 방법: RobustScaler\n",
      "열: NONLIVINGAPARTMENTS_MODE, 왜도: 29.13, 첨도: 1003.09, 스케일링 방법: RobustScaler\n",
      "열: NONLIVINGAREA_MODE, 왜도: 9.56, 첨도: 135.70, 스케일링 방법: RobustScaler\n",
      "열: APARTMENTS_MEDI, 왜도: 4.12, 첨도: 26.77, 스케일링 방법: RobustScaler\n",
      "열: BASEMENTAREA_MEDI, 왜도: 5.81, 첨도: 67.70, 스케일링 방법: RobustScaler\n",
      "열: YEARS_BEGINEXPLUATATION_MEDI, 왜도: -21.82, 첨도: 489.59, 스케일링 방법: RobustScaler\n",
      "열: YEARS_BUILD_MEDI, 왜도: -1.73, 첨도: 19.41, 스케일링 방법: RobustScaler\n",
      "열: COMMONAREA_MEDI, 왜도: 10.03, 첨도: 151.89, 스케일링 방법: RobustScaler\n",
      "열: ELEVATORS_MEDI, 왜도: 3.87, 첨도: 19.46, 스케일링 방법: RobustScaler\n",
      "열: ENTRANCES_MEDI, 왜도: 3.60, 첨도: 26.90, 스케일링 방법: RobustScaler\n",
      "열: FLOORSMAX_MEDI, 왜도: 2.32, 첨도: 8.42, 스케일링 방법: RobustScaler\n",
      "열: FLOORSMIN_MEDI, 왜도: 2.17, 첨도: 11.46, 스케일링 방법: RobustScaler\n",
      "열: LANDAREA_MEDI, 왜도: 7.16, 첨도: 86.68, 스케일링 방법: RobustScaler\n",
      "열: LIVINGAPARTMENTS_MEDI, 왜도: 5.85, 첨도: 57.29, 스케일링 방법: RobustScaler\n",
      "열: LIVINGAREA_MEDI, 왜도: 4.37, 첨도: 28.21, 스케일링 방법: RobustScaler\n",
      "열: NONLIVINGAPARTMENTS_MEDI, 왜도: 28.06, 첨도: 936.22, 스케일링 방법: RobustScaler\n",
      "열: NONLIVINGAREA_MEDI, 왜도: 9.57, 첨도: 137.09, 스케일링 방법: RobustScaler\n",
      "열: TOTALAREA_MODE, 왜도: 4.21, 첨도: 26.90, 스케일링 방법: RobustScaler\n",
      "열: EMERGENCYSTATE_MODE, 고유값: [0. 1.], 스케일링 제외 (바이너리 데이터)\n",
      "열: OBS_30_CNT_SOCIAL_CIRCLE, 왜도: 12.14, 첨도: 1426.49, 스케일링 방법: RobustScaler\n",
      "열: DEF_30_CNT_SOCIAL_CIRCLE, 왜도: 5.19, 첨도: 126.72, 스케일링 방법: RobustScaler\n",
      "열: OBS_60_CNT_SOCIAL_CIRCLE, 왜도: 12.08, 첨도: 1411.39, 스케일링 방법: RobustScaler\n",
      "열: DEF_60_CNT_SOCIAL_CIRCLE, 왜도: 5.29, 첨도: 86.86, 스케일링 방법: RobustScaler\n",
      "열: DAYS_LAST_PHONE_CHANGE, 왜도: -0.71, 첨도: -0.31, 스케일링 방법: RobustScaler\n",
      "열: FLAG_DOCUMENT_2, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_3, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_4, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_5, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_6, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_7, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_8, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_9, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_10, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_11, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_12, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_13, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_14, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_15, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_16, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_17, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_18, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_19, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_20, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FLAG_DOCUMENT_21, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: AMT_REQ_CREDIT_BUREAU_HOUR, 왜도: 15.64, 첨도: 294.62, 스케일링 방법: RobustScaler\n",
      "열: AMT_REQ_CREDIT_BUREAU_DAY, 왜도: 29.08, 첨도: 1331.74, 스케일링 방법: RobustScaler\n",
      "열: AMT_REQ_CREDIT_BUREAU_WEEK, 왜도: 10.01, 첨도: 192.73, 스케일링 방법: RobustScaler\n",
      "열: AMT_REQ_CREDIT_BUREAU_MON, 왜도: 8.37, 첨도: 103.98, 스케일링 방법: RobustScaler\n",
      "열: AMT_REQ_CREDIT_BUREAU_QRT, 왜도: 141.40, 첨도: 49065.79, 스케일링 방법: RobustScaler\n",
      "열: AMT_REQ_CREDIT_BUREAU_YEAR, 왜도: 1.45, 첨도: 2.77, 스케일링 방법: RobustScaler\n",
      "열: FONDKAPREMONT_MODE_org spec account, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: FONDKAPREMONT_MODE_reg oper account, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: FONDKAPREMONT_MODE_reg oper spec account, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: HOUSETYPE_MODE_ENCODED, 왜도: 11.92, 첨도: 146.21, 스케일링 방법: RobustScaler\n",
      "열: WALLSMATERIAL_MODE_ENCODED, 왜도: 1.36, 첨도: 9.32, 스케일링 방법: RobustScaler\n",
      "열: NAME_INCOME_TYPE_1.0, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: NAME_INCOME_TYPE_2.0, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: NAME_INCOME_TYPE_3.0, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n",
      "열: NAME_FAMILY_STATUS_2.0, 고유값: [1 0], 스케일링 제외 (바이너리 데이터)\n",
      "열: NAME_FAMILY_STATUS_3.0, 고유값: [0 1], 스케일링 제외 (바이너리 데이터)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'StandardScaler': 8, 'RobustScaler': 69, 'Skipped (Binary)': 49}"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler, RobustScaler\n",
    "import pandas as pd\n",
    "\n",
    "def automatic_scaling_with_skewness_kurtosis(df, binary_threshold=2, alpha=0.05):\n",
    "    scaled_columns = {}\n",
    "    scalers = {}\n",
    "    scaling_counts = {\"StandardScaler\": 0, \"RobustScaler\": 0, \"Skipped (Binary)\": 0}\n",
    "    \n",
    "    for col in df.select_dtypes(include=[\"float\", \"int\"]).columns:\n",
    "        unique_values = df[col].dropna().unique()\n",
    "        \n",
    "        # 바이너리 데이터 제외\n",
    "        if len(unique_values) <= binary_threshold:\n",
    "            print(f\"열: {col}, 고유값: {unique_values}, 스케일링 제외 (바이너리 데이터)\")\n",
    "            scaling_counts[\"Skipped (Binary)\"] += 1\n",
    "            scaled_columns[col] = df[col]  # 그대로 추가\n",
    "            continue\n",
    "        \n",
    "        # 왜도와 첨도 계산\n",
    "        skewness = df[col].dropna().skew()\n",
    "        kurtosis = df[col].dropna().kurtosis()\n",
    "        \n",
    "        # 스케일링 방법 선택\n",
    "        if abs(skewness) < 0.5 and abs(kurtosis) < 3:\n",
    "            scaler = StandardScaler()\n",
    "            scaling_method = 'StandardScaler'\n",
    "        else:\n",
    "            scaler = RobustScaler()\n",
    "            scaling_method = 'RobustScaler'\n",
    "        \n",
    "        # 스케일러를 적용하여 데이터 변환\n",
    "        scaled_values = scaler.fit_transform(df[[col]])\n",
    "        \n",
    "        # 스케일러 저장 및 카운트 증가\n",
    "        scalers[col] = scaler\n",
    "        scaling_counts[scaling_method] += 1\n",
    "        \n",
    "        # 변환된 값을 저장\n",
    "        scaled_columns[col] = scaled_values.flatten()\n",
    "        \n",
    "        print(f\"열: {col}, 왜도: {skewness:.2f}, 첨도: {kurtosis:.2f}, 스케일링 방법: {scaling_method}\")\n",
    "    \n",
    "    # DataFrame 생성: 모든 열을 한 번에 병합\n",
    "    scaled_df = pd.DataFrame(scaled_columns)\n",
    "    return scaled_df, scalers, scaling_counts\n",
    "\n",
    "scaled_df, scalers, scaling_counts = automatic_scaling_with_skewness_kurtosis(df_imputed)\n",
    "scaling_counts\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaled_df.to_csv(r\"C:\\Users\\Administrator\\Desktop\\scaled_by_skweness_and_kurtosis.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaled_df.drop(columns=\"Unnamed: 0\", inplace=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## VIF 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "VIF 계산 시작...\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Administrator\\AppData\\Local\\Temp\\ipykernel_28088\\3813816253.py:22: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation.\n",
      "  vif_df = pd.concat([vif_df, pd.DataFrame({\"Feature\": [feature_name], \"VIF\": [vif_value]})], ignore_index=True)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature: TARGET, VIF: 1.29\n",
      "Feature: NAME_CONTRACT_TYPE, VIF: 5.93\n",
      "Feature: CODE_GENDER, VIF: 4.01\n",
      "Feature: FLAG_OWN_CAR, VIF: 1.95\n",
      "Feature: FLAG_OWN_REALTY, VIF: 3.71\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Python312\\Lib\\site-packages\\statsmodels\\stats\\outliers_influence.py:197: RuntimeWarning: divide by zero encountered in scalar divide\n",
      "  vif = 1. / (1. - r_squared_i)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature: CNT_CHILDREN, VIF: inf\n",
      "Feature: AMT_INCOME_TOTAL, VIF: 1.07\n",
      "Feature: AMT_CREDIT, VIF: 44.26\n",
      "Feature: AMT_ANNUITY, VIF: 2.86\n",
      "Feature: AMT_GOODS_PRICE, VIF: 45.87\n",
      "Feature: NAME_TYPE_SUITE, VIF: 1.27\n",
      "Feature: NAME_EDUCATION_TYPE, VIF: 1.66\n",
      "Feature: NAME_HOUSING_TYPE, VIF: 9.88\n",
      "Feature: REGION_POPULATION_RELATIVE, VIF: 1.59\n",
      "Feature: DAYS_BIRTH, VIF: 2.50\n",
      "Feature: DAYS_EMPLOYED, VIF: 2569.47\n",
      "Feature: DAYS_REGISTRATION, VIF: 1.21\n",
      "Feature: DAYS_ID_PUBLISH, VIF: 1.15\n",
      "Feature: OWN_CAR_AGE, VIF: 1.10\n",
      "Feature: FLAG_MOBIL, VIF: 15061.27\n",
      "Feature: FLAG_EMP_PHONE, VIF: 11882.82\n",
      "Feature: FLAG_WORK_PHONE, VIF: 1.58\n",
      "Feature: FLAG_CONT_MOBILE, VIF: 569.65\n",
      "Feature: FLAG_PHONE, VIF: 1.61\n",
      "Feature: FLAG_EMAIL, VIF: 1.10\n",
      "Feature: OCCUPATION_TYPE, VIF: 1.31\n",
      "Feature: CNT_FAM_MEMBERS, VIF: inf\n",
      "Feature: REGION_RATING_CLIENT, VIF: 11.03\n",
      "Feature: REGION_RATING_CLIENT_W_CITY, VIF: 10.90\n",
      "Feature: WEEKDAY_APPR_PROCESS_START, VIF: 1.21\n",
      "Feature: HOUR_APPR_PROCESS_START, VIF: 1.15\n",
      "Feature: REG_REGION_NOT_LIVE_REGION, VIF: 2.47\n",
      "Feature: REG_REGION_NOT_WORK_REGION, VIF: 9.56\n",
      "Feature: LIVE_REGION_NOT_WORK_REGION, VIF: 7.67\n",
      "Feature: REG_CITY_NOT_LIVE_CITY, VIF: 2.65\n",
      "Feature: REG_CITY_NOT_WORK_CITY, VIF: 9.81\n",
      "Feature: LIVE_CITY_NOT_WORK_CITY, VIF: 7.36\n",
      "Feature: ORGANIZATION_TYPE, VIF: 1.28\n",
      "Feature: EXT_SOURCE_1, VIF: 1.35\n",
      "Feature: EXT_SOURCE_2, VIF: 1.32\n",
      "Feature: EXT_SOURCE_3, VIF: 1.15\n",
      "Feature: APARTMENTS_AVG, VIF: 271.30\n",
      "Feature: BASEMENTAREA_AVG, VIF: 106.57\n",
      "Feature: YEARS_BEGINEXPLUATATION_AVG, VIF: 108.03\n",
      "Feature: YEARS_BUILD_AVG, VIF: 380.07\n",
      "Feature: COMMONAREA_AVG, VIF: 148.60\n",
      "Feature: ELEVATORS_AVG, VIF: 324.71\n",
      "Feature: ENTRANCES_AVG, VIF: 261.63\n",
      "Feature: FLOORSMAX_AVG, VIF: 281.07\n",
      "Feature: FLOORSMIN_AVG, VIF: 234.08\n",
      "Feature: LANDAREA_AVG, VIF: 64.33\n",
      "Feature: LIVINGAPARTMENTS_AVG, VIF: 171.52\n",
      "Feature: LIVINGAREA_AVG, VIF: 212.75\n",
      "Feature: NONLIVINGAPARTMENTS_AVG, VIF: 63.39\n",
      "Feature: NONLIVINGAREA_AVG, VIF: 65.86\n",
      "Feature: APARTMENTS_MODE, VIF: 80.43\n",
      "Feature: BASEMENTAREA_MODE, VIF: 40.58\n",
      "Feature: YEARS_BEGINEXPLUATATION_MODE, VIF: 19.01\n",
      "Feature: YEARS_BUILD_MODE, VIF: 60.87\n",
      "Feature: COMMONAREA_MODE, VIF: 37.67\n",
      "Feature: ELEVATORS_MODE, VIF: 98.24\n",
      "Feature: ENTRANCES_MODE, VIF: 52.99\n",
      "Feature: FLOORSMAX_MODE, VIF: 75.33\n",
      "Feature: FLOORSMIN_MODE, VIF: 64.27\n",
      "Feature: LANDAREA_MODE, VIF: 35.90\n",
      "Feature: LIVINGAPARTMENTS_MODE, VIF: 41.92\n",
      "Feature: LIVINGAREA_MODE, VIF: 57.65\n",
      "Feature: NONLIVINGAPARTMENTS_MODE, VIF: 26.74\n",
      "Feature: NONLIVINGAREA_MODE, VIF: 29.34\n",
      "Feature: APARTMENTS_MEDI, VIF: 309.20\n",
      "Feature: BASEMENTAREA_MEDI, VIF: 126.62\n",
      "Feature: YEARS_BEGINEXPLUATATION_MEDI, VIF: 83.54\n",
      "Feature: YEARS_BUILD_MEDI, VIF: 378.81\n",
      "Feature: COMMONAREA_MEDI, VIF: 169.14\n",
      "Feature: ELEVATORS_MEDI, VIF: 402.34\n",
      "Feature: ENTRANCES_MEDI, VIF: 296.35\n",
      "Feature: FLOORSMAX_MEDI, VIF: 333.52\n",
      "Feature: FLOORSMIN_MEDI, VIF: 286.01\n",
      "Feature: LANDAREA_MEDI, VIF: 92.33\n",
      "Feature: LIVINGAPARTMENTS_MEDI, VIF: 193.05\n",
      "Feature: LIVINGAREA_MEDI, VIF: 240.04\n",
      "Feature: NONLIVINGAPARTMENTS_MEDI, VIF: 88.28\n",
      "Feature: NONLIVINGAREA_MEDI, VIF: 89.33\n",
      "Feature: TOTALAREA_MODE, VIF: 8.94\n",
      "Feature: EMERGENCYSTATE_MODE, VIF: 1.07\n",
      "Feature: OBS_30_CNT_SOCIAL_CIRCLE, VIF: 448.34\n",
      "Feature: DEF_30_CNT_SOCIAL_CIRCLE, VIF: 4.48\n",
      "Feature: OBS_60_CNT_SOCIAL_CIRCLE, VIF: 448.32\n",
      "Feature: DEF_60_CNT_SOCIAL_CIRCLE, VIF: 4.17\n",
      "Feature: DAYS_LAST_PHONE_CHANGE, VIF: 1.18\n",
      "Feature: FLAG_DOCUMENT_2, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_3, VIF: 42.34\n",
      "Feature: FLAG_DOCUMENT_4, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_5, VIF: 1.65\n",
      "Feature: FLAG_DOCUMENT_6, VIF: 6.82\n",
      "Feature: FLAG_DOCUMENT_7, VIF: 1.01\n",
      "Feature: FLAG_DOCUMENT_8, VIF: 5.77\n",
      "Feature: FLAG_DOCUMENT_9, VIF: 1.21\n",
      "Feature: FLAG_DOCUMENT_10, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_11, VIF: 1.16\n",
      "Feature: FLAG_DOCUMENT_12, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_13, VIF: 1.04\n",
      "Feature: FLAG_DOCUMENT_14, VIF: 1.02\n",
      "Feature: FLAG_DOCUMENT_15, VIF: 1.01\n",
      "Feature: FLAG_DOCUMENT_16, VIF: 1.05\n",
      "Feature: FLAG_DOCUMENT_17, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_18, VIF: 1.04\n",
      "Feature: FLAG_DOCUMENT_19, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_20, VIF: 1.02\n",
      "Feature: FLAG_DOCUMENT_21, VIF: 1.01\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_HOUR, VIF: 1.07\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_DAY, VIF: 1.12\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_WEEK, VIF: 1.08\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_MON, VIF: 1.09\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_QRT, VIF: 1.11\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_YEAR, VIF: 1.28\n",
      "Feature: FONDKAPREMONT_MODE_org spec account, VIF: 1.09\n",
      "Feature: FONDKAPREMONT_MODE_reg oper account, VIF: 1.68\n",
      "Feature: FONDKAPREMONT_MODE_reg oper spec account, VIF: 1.12\n",
      "Feature: HOUSETYPE_MODE_ENCODED, VIF: 1.05\n",
      "Feature: WALLSMATERIAL_MODE_ENCODED, VIF: 1.39\n",
      "Feature: NAME_INCOME_TYPE_1.0, VIF: 1.58\n",
      "Feature: NAME_INCOME_TYPE_2.0, VIF: 4.36\n",
      "Feature: NAME_INCOME_TYPE_3.0, VIF: 1.01\n",
      "Feature: NAME_FAMILY_STATUS_2.0, VIF: inf\n",
      "Feature: NAME_FAMILY_STATUS_3.0, VIF: inf\n",
      "\n",
      "VIF 값이 임계값을 초과하는 변수들:\n",
      "                      Feature          VIF\n",
      "5                CNT_CHILDREN          inf\n",
      "7                  AMT_CREDIT    44.260245\n",
      "9             AMT_GOODS_PRICE    45.868002\n",
      "12          NAME_HOUSING_TYPE     9.883604\n",
      "15              DAYS_EMPLOYED  2569.466227\n",
      "..                        ...          ...\n",
      "85   OBS_30_CNT_SOCIAL_CIRCLE   448.340595\n",
      "87   OBS_60_CNT_SOCIAL_CIRCLE   448.317586\n",
      "91            FLAG_DOCUMENT_3    42.342236\n",
      "124    NAME_FAMILY_STATUS_2.0          inf\n",
      "125    NAME_FAMILY_STATUS_3.0          inf\n",
      "\n",
      "[61 rows x 2 columns]\n",
      "\n",
      "VIF가 높은 변수들 리스트:\n",
      "['CNT_CHILDREN', 'AMT_CREDIT', 'AMT_GOODS_PRICE', 'NAME_HOUSING_TYPE', 'DAYS_EMPLOYED', 'FLAG_MOBIL', 'FLAG_EMP_PHONE', 'FLAG_CONT_MOBILE', 'CNT_FAM_MEMBERS', 'REGION_RATING_CLIENT', 'REGION_RATING_CLIENT_W_CITY', 'REG_REGION_NOT_WORK_REGION', 'REG_CITY_NOT_WORK_CITY', 'APARTMENTS_AVG', 'BASEMENTAREA_AVG', 'YEARS_BEGINEXPLUATATION_AVG', 'YEARS_BUILD_AVG', 'COMMONAREA_AVG', 'ELEVATORS_AVG', 'ENTRANCES_AVG', 'FLOORSMAX_AVG', 'FLOORSMIN_AVG', 'LANDAREA_AVG', 'LIVINGAPARTMENTS_AVG', 'LIVINGAREA_AVG', 'NONLIVINGAPARTMENTS_AVG', 'NONLIVINGAREA_AVG', 'APARTMENTS_MODE', 'BASEMENTAREA_MODE', 'YEARS_BEGINEXPLUATATION_MODE', 'YEARS_BUILD_MODE', 'COMMONAREA_MODE', 'ELEVATORS_MODE', 'ENTRANCES_MODE', 'FLOORSMAX_MODE', 'FLOORSMIN_MODE', 'LANDAREA_MODE', 'LIVINGAPARTMENTS_MODE', 'LIVINGAREA_MODE', 'NONLIVINGAPARTMENTS_MODE', 'NONLIVINGAREA_MODE', 'APARTMENTS_MEDI', 'BASEMENTAREA_MEDI', 'YEARS_BEGINEXPLUATATION_MEDI', 'YEARS_BUILD_MEDI', 'COMMONAREA_MEDI', 'ELEVATORS_MEDI', 'ENTRANCES_MEDI', 'FLOORSMAX_MEDI', 'FLOORSMIN_MEDI', 'LANDAREA_MEDI', 'LIVINGAPARTMENTS_MEDI', 'LIVINGAREA_MEDI', 'NONLIVINGAPARTMENTS_MEDI', 'NONLIVINGAREA_MEDI', 'TOTALAREA_MODE', 'OBS_30_CNT_SOCIAL_CIRCLE', 'OBS_60_CNT_SOCIAL_CIRCLE', 'FLAG_DOCUMENT_3', 'NAME_FAMILY_STATUS_2.0', 'NAME_FAMILY_STATUS_3.0']\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
    "from statsmodels.tools.tools import add_constant\n",
    "\n",
    "def check_vif_live(df, thresh=100):\n",
    "    \"\"\"\n",
    "    VIF 값을 계산하고 한 번 계산할 때마다 출력하는 함수.\n",
    "    Parameters:\n",
    "        - df (pd.DataFrame): VIF 계산 대상 데이터프레임\n",
    "        - thresh (float): VIF 임계값 (기본값: 8)\n",
    "    Returns:\n",
    "        - pd.DataFrame: VIF가 임계값을 초과하는 변수들 및 VIF 값\n",
    "    \"\"\"\n",
    "    vif_df = pd.DataFrame(columns=[\"Feature\", \"VIF\"])\n",
    "    \n",
    "    print(\"VIF 계산 시작...\\n\")\n",
    "    \n",
    "    # VIF 계산\n",
    "    for i in range(df.shape[1]):\n",
    "        feature_name = df.columns[i]\n",
    "        vif_value = variance_inflation_factor(df.values, i)\n",
    "        vif_df = pd.concat([vif_df, pd.DataFrame({\"Feature\": [feature_name], \"VIF\": [vif_value]})], ignore_index=True)\n",
    "        \n",
    "        # 계산한 VIF 바로 출력\n",
    "        print(f\"Feature: {feature_name}, VIF: {vif_value:.2f}\")\n",
    "    \n",
    "    # VIF가 임계값을 초과하는 변수 필터링\n",
    "    high_vif = vif_df[vif_df[\"VIF\"] > thresh]\n",
    "    \n",
    "    print(\"\\nVIF 값이 임계값을 초과하는 변수들:\")\n",
    "    print(high_vif)\n",
    "\n",
    "    return high_vif\n",
    "\n",
    "# 예제: scaled_df에 VIF 확인\n",
    "high_vif_features = check_vif_live(scaled_df)\n",
    "\n",
    "# VIF가 높은 변수 리스트 출력\n",
    "print(\"\\nVIF가 높은 변수들 리스트:\")\n",
    "print(high_vif_features[\"Feature\"].tolist())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "high_vif = ['CNT_CHILDREN', 'AMT_CREDIT', 'AMT_GOODS_PRICE', 'NAME_HOUSING_TYPE', 'DAYS_EMPLOYED', 'FLAG_MOBIL', 'FLAG_EMP_PHONE', 'FLAG_CONT_MOBILE', 'CNT_FAM_MEMBERS', 'REGION_RATING_CLIENT', 'REGION_RATING_CLIENT_W_CITY', 'REG_REGION_NOT_WORK_REGION', 'REG_CITY_NOT_WORK_CITY', 'APARTMENTS_AVG', 'BASEMENTAREA_AVG', 'YEARS_BEGINEXPLUATATION_AVG', 'YEARS_BUILD_AVG', 'COMMONAREA_AVG', 'ELEVATORS_AVG', 'ENTRANCES_AVG', 'FLOORSMAX_AVG', 'FLOORSMIN_AVG', 'LANDAREA_AVG', 'LIVINGAPARTMENTS_AVG', 'LIVINGAREA_AVG', 'NONLIVINGAPARTMENTS_AVG', 'NONLIVINGAREA_AVG', 'APARTMENTS_MODE', 'BASEMENTAREA_MODE', 'YEARS_BEGINEXPLUATATION_MODE', 'YEARS_BUILD_MODE', 'COMMONAREA_MODE', 'ELEVATORS_MODE', 'ENTRANCES_MODE', 'FLOORSMAX_MODE', 'FLOORSMIN_MODE', 'LANDAREA_MODE', 'LIVINGAPARTMENTS_MODE', 'LIVINGAREA_MODE', 'NONLIVINGAPARTMENTS_MODE', 'NONLIVINGAREA_MODE', 'APARTMENTS_MEDI', 'BASEMENTAREA_MEDI', 'YEARS_BEGINEXPLUATATION_MEDI', 'YEARS_BUILD_MEDI', 'COMMONAREA_MEDI', 'ELEVATORS_MEDI', 'ENTRANCES_MEDI', 'FLOORSMAX_MEDI', 'FLOORSMIN_MEDI', 'LANDAREA_MEDI', 'LIVINGAPARTMENTS_MEDI', 'LIVINGAREA_MEDI', 'NONLIVINGAPARTMENTS_MEDI', 'NONLIVINGAREA_MEDI', 'TOTALAREA_MODE', 'OBS_30_CNT_SOCIAL_CIRCLE', 'OBS_60_CNT_SOCIAL_CIRCLE', 'FLAG_DOCUMENT_3', 'NAME_FAMILY_STATUS_2.0', 'NAME_FAMILY_STATUS_3.0']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "VIF 계산 시작...\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Python312\\Lib\\site-packages\\statsmodels\\stats\\outliers_influence.py:197: RuntimeWarning: divide by zero encountered in scalar divide\n",
      "  vif = 1. / (1. - r_squared_i)\n",
      "C:\\Users\\Administrator\\AppData\\Local\\Temp\\ipykernel_28088\\170157159.py:25: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation.\n",
      "  vif_df = pd.concat([vif_df, pd.DataFrame({\"Feature\": [feature_name], \"VIF\": [vif_value]})], ignore_index=True)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature: CNT_CHILDREN, VIF: inf (계산됨)\n",
      "Feature: AMT_CREDIT, VIF: 44.26 (계산됨)\n",
      "Feature: AMT_GOODS_PRICE, VIF: 45.87 (계산됨)\n",
      "Feature: NAME_HOUSING_TYPE, VIF: 9.88 (계산됨)\n",
      "Feature: DAYS_EMPLOYED, VIF: 2569.47 (계산됨)\n",
      "Feature: FLAG_MOBIL, VIF: 15061.27 (계산됨)\n",
      "Feature: FLAG_EMP_PHONE, VIF: 11882.82 (계산됨)\n",
      "Feature: FLAG_CONT_MOBILE, VIF: 569.65 (계산됨)\n",
      "Feature: CNT_FAM_MEMBERS, VIF: inf (계산됨)\n",
      "Feature: REGION_RATING_CLIENT, VIF: 11.03 (계산됨)\n",
      "Feature: REGION_RATING_CLIENT_W_CITY, VIF: 10.90 (계산됨)\n",
      "Feature: REG_REGION_NOT_WORK_REGION, VIF: 9.56 (계산됨)\n",
      "Feature: REG_CITY_NOT_WORK_CITY, VIF: 9.81 (계산됨)\n",
      "Feature: APARTMENTS_AVG, VIF: 271.30 (계산됨)\n",
      "Feature: BASEMENTAREA_AVG, VIF: 106.57 (계산됨)\n",
      "Feature: YEARS_BEGINEXPLUATATION_AVG, VIF: 108.03 (계산됨)\n",
      "Feature: YEARS_BUILD_AVG, VIF: 380.07 (계산됨)\n",
      "Feature: COMMONAREA_AVG, VIF: 148.60 (계산됨)\n",
      "Feature: ELEVATORS_AVG, VIF: 324.71 (계산됨)\n",
      "Feature: ENTRANCES_AVG, VIF: 261.63 (계산됨)\n",
      "Feature: FLOORSMAX_AVG, VIF: 281.07 (계산됨)\n",
      "Feature: FLOORSMIN_AVG, VIF: 234.08 (계산됨)\n",
      "Feature: LANDAREA_AVG, VIF: 64.33 (계산됨)\n",
      "Feature: LIVINGAPARTMENTS_AVG, VIF: 171.52 (계산됨)\n",
      "Feature: LIVINGAREA_AVG, VIF: 212.75 (계산됨)\n",
      "Feature: NONLIVINGAPARTMENTS_AVG, VIF: 63.39 (계산됨)\n",
      "Feature: NONLIVINGAREA_AVG, VIF: 65.86 (계산됨)\n",
      "Feature: APARTMENTS_MODE, VIF: 80.43 (계산됨)\n",
      "Feature: BASEMENTAREA_MODE, VIF: 40.58 (계산됨)\n",
      "Feature: YEARS_BEGINEXPLUATATION_MODE, VIF: 19.01 (계산됨)\n",
      "Feature: YEARS_BUILD_MODE, VIF: 60.87 (계산됨)\n",
      "Feature: COMMONAREA_MODE, VIF: 37.67 (계산됨)\n",
      "Feature: ELEVATORS_MODE, VIF: 98.24 (계산됨)\n",
      "Feature: ENTRANCES_MODE, VIF: 52.99 (계산됨)\n",
      "Feature: FLOORSMAX_MODE, VIF: 75.33 (계산됨)\n",
      "Feature: FLOORSMIN_MODE, VIF: 64.27 (계산됨)\n",
      "Feature: LANDAREA_MODE, VIF: 35.90 (계산됨)\n",
      "Feature: LIVINGAPARTMENTS_MODE, VIF: 41.92 (계산됨)\n",
      "Feature: LIVINGAREA_MODE, VIF: 57.65 (계산됨)\n",
      "Feature: NONLIVINGAPARTMENTS_MODE, VIF: 26.74 (계산됨)\n",
      "Feature: NONLIVINGAREA_MODE, VIF: 29.34 (계산됨)\n",
      "Feature: APARTMENTS_MEDI, VIF: 309.20 (계산됨)\n",
      "Feature: BASEMENTAREA_MEDI, VIF: 126.62 (계산됨)\n",
      "Feature: YEARS_BEGINEXPLUATATION_MEDI, VIF: 83.54 (계산됨)\n",
      "Feature: YEARS_BUILD_MEDI, VIF: 378.81 (계산됨)\n",
      "Feature: COMMONAREA_MEDI, VIF: 169.14 (계산됨)\n",
      "Feature: ELEVATORS_MEDI, VIF: 402.34 (계산됨)\n",
      "Feature: ENTRANCES_MEDI, VIF: 296.35 (계산됨)\n",
      "Feature: FLOORSMAX_MEDI, VIF: 333.52 (계산됨)\n",
      "Feature: FLOORSMIN_MEDI, VIF: 286.01 (계산됨)\n",
      "Feature: LANDAREA_MEDI, VIF: 92.33 (계산됨)\n",
      "Feature: LIVINGAPARTMENTS_MEDI, VIF: 193.05 (계산됨)\n",
      "Feature: LIVINGAREA_MEDI, VIF: 240.04 (계산됨)\n",
      "Feature: NONLIVINGAPARTMENTS_MEDI, VIF: 88.28 (계산됨)\n",
      "Feature: NONLIVINGAREA_MEDI, VIF: 89.33 (계산됨)\n",
      "Feature: TOTALAREA_MODE, VIF: 8.94 (계산됨)\n",
      "Feature: OBS_30_CNT_SOCIAL_CIRCLE, VIF: 448.34 (계산됨)\n",
      "Feature: OBS_60_CNT_SOCIAL_CIRCLE, VIF: 448.32 (계산됨)\n",
      "Feature: FLAG_DOCUMENT_3, VIF: 42.34 (계산됨)\n",
      "Feature: NAME_FAMILY_STATUS_2.0, VIF: inf (계산됨)\n",
      "Feature: NAME_FAMILY_STATUS_3.0, VIF: inf (계산됨)\n",
      "\n",
      "VIF 계산 완료!\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
    "\n",
    "def calculate_vif_for_selected(df, high_vif, thresh=100):\n",
    "    \"\"\"\n",
    "    전체 데이터프레임에서 high_vif 리스트에 포함된 변수만 VIF 계산.\n",
    "    Parameters:\n",
    "        - df (pd.DataFrame): 전체 데이터프레임\n",
    "        - high_vif (list): VIF를 계산할 변수 이름 리스트\n",
    "        - thresh (float): VIF 임계값 (기본값: 100)\n",
    "    Returns:\n",
    "        - pd.DataFrame: high_vif 변수들의 VIF 값 데이터프레임\n",
    "        - pd.DataFrame: high_vif 변수들 중 VIF > 100 변수들의 데이터프레임\n",
    "    \"\"\"\n",
    "    vif_df = pd.DataFrame(columns=[\"Feature\", \"VIF\"])\n",
    "\n",
    "    print(\"VIF 계산 시작...\\n\")\n",
    "\n",
    "    # high_vif 리스트에 포함된 변수만 계산\n",
    "    for feature_name in high_vif:\n",
    "        if feature_name in df.columns:\n",
    "            # 전체 데이터프레임에서 high_vif 변수에 대해 VIF 계산\n",
    "            index = list(df.columns).index(feature_name)  # 해당 변수의 인덱스\n",
    "            vif_value = variance_inflation_factor(df.values, index)\n",
    "            vif_df = pd.concat([vif_df, pd.DataFrame({\"Feature\": [feature_name], \"VIF\": [vif_value]})], ignore_index=True)\n",
    "            print(f\"Feature: {feature_name}, VIF: {vif_value:.2f} (계산됨)\")\n",
    "        else:\n",
    "            print(f\"Feature: {feature_name} (데이터프레임에 존재하지 않음, 스킵됨)\")\n",
    "\n",
    "    print(\"\\nVIF 계산 완료!\\n\")\n",
    "\n",
    "    # VIF > thresh 필터링\n",
    "    super_high_vif = vif_df[vif_df[\"VIF\"] > thresh]\n",
    "\n",
    "    return vif_df, super_high_vif\n",
    "vif_results, super_high_vif = calculate_vif_for_selected(scaled_df, high_vif)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "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>Feature</th>\n",
       "      <th>VIF</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CNT_CHILDREN</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>DAYS_EMPLOYED</td>\n",
       "      <td>2.569466e+03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>FLAG_MOBIL</td>\n",
       "      <td>1.506127e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>FLAG_EMP_PHONE</td>\n",
       "      <td>1.188282e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>FLAG_CONT_MOBILE</td>\n",
       "      <td>5.696474e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>CNT_FAM_MEMBERS</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>APARTMENTS_AVG</td>\n",
       "      <td>2.712962e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>BASEMENTAREA_AVG</td>\n",
       "      <td>1.065663e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>YEARS_BEGINEXPLUATATION_AVG</td>\n",
       "      <td>1.080287e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>YEARS_BUILD_AVG</td>\n",
       "      <td>3.800748e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>COMMONAREA_AVG</td>\n",
       "      <td>1.485991e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>ELEVATORS_AVG</td>\n",
       "      <td>3.247097e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>ENTRANCES_AVG</td>\n",
       "      <td>2.616325e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>FLOORSMAX_AVG</td>\n",
       "      <td>2.810691e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>FLOORSMIN_AVG</td>\n",
       "      <td>2.340827e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>LIVINGAPARTMENTS_AVG</td>\n",
       "      <td>1.715198e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>LIVINGAREA_AVG</td>\n",
       "      <td>2.127548e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>APARTMENTS_MEDI</td>\n",
       "      <td>3.092024e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>BASEMENTAREA_MEDI</td>\n",
       "      <td>1.266199e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>YEARS_BUILD_MEDI</td>\n",
       "      <td>3.788091e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>COMMONAREA_MEDI</td>\n",
       "      <td>1.691386e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>ELEVATORS_MEDI</td>\n",
       "      <td>4.023398e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>ENTRANCES_MEDI</td>\n",
       "      <td>2.963475e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>FLOORSMAX_MEDI</td>\n",
       "      <td>3.335227e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>FLOORSMIN_MEDI</td>\n",
       "      <td>2.860149e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>LIVINGAPARTMENTS_MEDI</td>\n",
       "      <td>1.930464e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>LIVINGAREA_MEDI</td>\n",
       "      <td>2.400408e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>OBS_30_CNT_SOCIAL_CIRCLE</td>\n",
       "      <td>4.483406e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>OBS_60_CNT_SOCIAL_CIRCLE</td>\n",
       "      <td>4.483176e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>NAME_FAMILY_STATUS_2.0</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>NAME_FAMILY_STATUS_3.0</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        Feature           VIF\n",
       "0                  CNT_CHILDREN           inf\n",
       "4                 DAYS_EMPLOYED  2.569466e+03\n",
       "5                    FLAG_MOBIL  1.506127e+04\n",
       "6                FLAG_EMP_PHONE  1.188282e+04\n",
       "7              FLAG_CONT_MOBILE  5.696474e+02\n",
       "8               CNT_FAM_MEMBERS           inf\n",
       "13               APARTMENTS_AVG  2.712962e+02\n",
       "14             BASEMENTAREA_AVG  1.065663e+02\n",
       "15  YEARS_BEGINEXPLUATATION_AVG  1.080287e+02\n",
       "16              YEARS_BUILD_AVG  3.800748e+02\n",
       "17               COMMONAREA_AVG  1.485991e+02\n",
       "18                ELEVATORS_AVG  3.247097e+02\n",
       "19                ENTRANCES_AVG  2.616325e+02\n",
       "20                FLOORSMAX_AVG  2.810691e+02\n",
       "21                FLOORSMIN_AVG  2.340827e+02\n",
       "23         LIVINGAPARTMENTS_AVG  1.715198e+02\n",
       "24               LIVINGAREA_AVG  2.127548e+02\n",
       "41              APARTMENTS_MEDI  3.092024e+02\n",
       "42            BASEMENTAREA_MEDI  1.266199e+02\n",
       "44             YEARS_BUILD_MEDI  3.788091e+02\n",
       "45              COMMONAREA_MEDI  1.691386e+02\n",
       "46               ELEVATORS_MEDI  4.023398e+02\n",
       "47               ENTRANCES_MEDI  2.963475e+02\n",
       "48               FLOORSMAX_MEDI  3.335227e+02\n",
       "49               FLOORSMIN_MEDI  2.860149e+02\n",
       "51        LIVINGAPARTMENTS_MEDI  1.930464e+02\n",
       "52              LIVINGAREA_MEDI  2.400408e+02\n",
       "56     OBS_30_CNT_SOCIAL_CIRCLE  4.483406e+02\n",
       "57     OBS_60_CNT_SOCIAL_CIRCLE  4.483176e+02\n",
       "59       NAME_FAMILY_STATUS_2.0           inf\n",
       "60       NAME_FAMILY_STATUS_3.0           inf"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "super_high_vif"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "확인결과 MEDI의 값들과 AVG의 값들이 전부 vif지수가 강하게 나왔다. 또한 중앙, 최빈, 평균값은 사실상 비슷한 것을 의미하기 때문에 vif지수가 덜 높은 최빈값만을 남기고 나머지는 삭제한다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MEDI, AVG 삭제 후 다시 VIF 지수 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "삭제된 열: ['APARTMENTS_AVG', 'BASEMENTAREA_AVG', 'YEARS_BEGINEXPLUATATION_AVG', 'YEARS_BUILD_AVG', 'COMMONAREA_AVG', 'ELEVATORS_AVG', 'ENTRANCES_AVG', 'FLOORSMAX_AVG', 'FLOORSMIN_AVG', 'LANDAREA_AVG', 'LIVINGAPARTMENTS_AVG', 'LIVINGAREA_AVG', 'NONLIVINGAPARTMENTS_AVG', 'NONLIVINGAREA_AVG', 'APARTMENTS_MEDI', 'BASEMENTAREA_MEDI', 'YEARS_BEGINEXPLUATATION_MEDI', 'YEARS_BUILD_MEDI', 'COMMONAREA_MEDI', 'ELEVATORS_MEDI', 'ENTRANCES_MEDI', 'FLOORSMAX_MEDI', 'FLOORSMIN_MEDI', 'LANDAREA_MEDI', 'LIVINGAPARTMENTS_MEDI', 'LIVINGAREA_MEDI', 'NONLIVINGAPARTMENTS_MEDI', 'NONLIVINGAREA_MEDI']\n",
      "새로운 데이터프레임 크기: (307507, 98)\n"
     ]
    }
   ],
   "source": [
    "# '_MEDI' 또는 '_AVG'로 끝나는 열을 삭제\n",
    "columns_to_drop = [col for col in scaled_df.columns if col.endswith('_MEDI') or col.endswith('_AVG')]\n",
    "df_drop_avg_medi = scaled_df.drop(columns=columns_to_drop)\n",
    "\n",
    "# 결과 확인\n",
    "print(f\"삭제된 열: {columns_to_drop}\")\n",
    "print(f\"새로운 데이터프레임 크기: {df_drop_avg_medi.shape}\")\n",
    "df_drop_avg_medi_data = df_drop_avg_medi.drop(columns=\"TARGET\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "VIF 계산 시작...\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Administrator\\AppData\\Local\\Temp\\ipykernel_28088\\301706835.py:22: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation.\n",
      "  vif_df = pd.concat([vif_df, pd.DataFrame({\"Feature\": [feature_name], \"VIF\": [vif_value]})], ignore_index=True)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature: NAME_CONTRACT_TYPE, VIF: 5.93\n",
      "Feature: CODE_GENDER, VIF: 4.01\n",
      "Feature: FLAG_OWN_CAR, VIF: 1.95\n",
      "Feature: FLAG_OWN_REALTY, VIF: 3.71\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Python312\\Lib\\site-packages\\statsmodels\\stats\\outliers_influence.py:197: RuntimeWarning: divide by zero encountered in scalar divide\n",
      "  vif = 1. / (1. - r_squared_i)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature: CNT_CHILDREN, VIF: inf\n",
      "Feature: AMT_INCOME_TOTAL, VIF: 1.07\n",
      "Feature: AMT_CREDIT, VIF: 44.21\n",
      "Feature: AMT_ANNUITY, VIF: 2.85\n",
      "Feature: AMT_GOODS_PRICE, VIF: 45.81\n",
      "Feature: NAME_TYPE_SUITE, VIF: 1.27\n",
      "Feature: NAME_EDUCATION_TYPE, VIF: 1.66\n",
      "Feature: NAME_HOUSING_TYPE, VIF: 9.88\n",
      "Feature: REGION_POPULATION_RELATIVE, VIF: 1.57\n",
      "Feature: DAYS_BIRTH, VIF: 2.49\n",
      "Feature: DAYS_EMPLOYED, VIF: 2568.54\n",
      "Feature: DAYS_REGISTRATION, VIF: 1.20\n",
      "Feature: DAYS_ID_PUBLISH, VIF: 1.15\n",
      "Feature: OWN_CAR_AGE, VIF: 1.09\n",
      "Feature: FLAG_MOBIL, VIF: 15054.19\n",
      "Feature: FLAG_EMP_PHONE, VIF: 11878.13\n",
      "Feature: FLAG_WORK_PHONE, VIF: 1.58\n",
      "Feature: FLAG_CONT_MOBILE, VIF: 569.56\n",
      "Feature: FLAG_PHONE, VIF: 1.61\n",
      "Feature: FLAG_EMAIL, VIF: 1.10\n",
      "Feature: OCCUPATION_TYPE, VIF: 1.31\n",
      "Feature: CNT_FAM_MEMBERS, VIF: inf\n",
      "Feature: REGION_RATING_CLIENT, VIF: 11.02\n",
      "Feature: REGION_RATING_CLIENT_W_CITY, VIF: 10.89\n",
      "Feature: WEEKDAY_APPR_PROCESS_START, VIF: 1.21\n",
      "Feature: HOUR_APPR_PROCESS_START, VIF: 1.15\n",
      "Feature: REG_REGION_NOT_LIVE_REGION, VIF: 2.47\n",
      "Feature: REG_REGION_NOT_WORK_REGION, VIF: 9.56\n",
      "Feature: LIVE_REGION_NOT_WORK_REGION, VIF: 7.67\n",
      "Feature: REG_CITY_NOT_LIVE_CITY, VIF: 2.65\n",
      "Feature: REG_CITY_NOT_WORK_CITY, VIF: 9.80\n",
      "Feature: LIVE_CITY_NOT_WORK_CITY, VIF: 7.36\n",
      "Feature: ORGANIZATION_TYPE, VIF: 1.28\n",
      "Feature: EXT_SOURCE_1, VIF: 1.29\n",
      "Feature: EXT_SOURCE_2, VIF: 1.30\n",
      "Feature: EXT_SOURCE_3, VIF: 1.12\n",
      "Feature: APARTMENTS_MODE, VIF: 7.81\n",
      "Feature: BASEMENTAREA_MODE, VIF: 2.11\n",
      "Feature: YEARS_BEGINEXPLUATATION_MODE, VIF: 1.04\n",
      "Feature: YEARS_BUILD_MODE, VIF: 1.36\n",
      "Feature: COMMONAREA_MODE, VIF: 1.53\n",
      "Feature: ELEVATORS_MODE, VIF: 5.50\n",
      "Feature: ENTRANCES_MODE, VIF: 2.73\n",
      "Feature: FLOORSMAX_MODE, VIF: 3.88\n",
      "Feature: FLOORSMIN_MODE, VIF: 1.71\n",
      "Feature: LANDAREA_MODE, VIF: 1.51\n",
      "Feature: LIVINGAPARTMENTS_MODE, VIF: 3.15\n",
      "Feature: LIVINGAREA_MODE, VIF: 8.86\n",
      "Feature: NONLIVINGAPARTMENTS_MODE, VIF: 1.08\n",
      "Feature: NONLIVINGAREA_MODE, VIF: 1.29\n",
      "Feature: TOTALAREA_MODE, VIF: 6.47\n",
      "Feature: EMERGENCYSTATE_MODE, VIF: 1.07\n",
      "Feature: OBS_30_CNT_SOCIAL_CIRCLE, VIF: 448.32\n",
      "Feature: DEF_30_CNT_SOCIAL_CIRCLE, VIF: 4.48\n",
      "Feature: OBS_60_CNT_SOCIAL_CIRCLE, VIF: 448.29\n",
      "Feature: DEF_60_CNT_SOCIAL_CIRCLE, VIF: 4.17\n",
      "Feature: DAYS_LAST_PHONE_CHANGE, VIF: 1.18\n",
      "Feature: FLAG_DOCUMENT_2, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_3, VIF: 42.33\n",
      "Feature: FLAG_DOCUMENT_4, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_5, VIF: 1.65\n",
      "Feature: FLAG_DOCUMENT_6, VIF: 6.82\n",
      "Feature: FLAG_DOCUMENT_7, VIF: 1.01\n",
      "Feature: FLAG_DOCUMENT_8, VIF: 5.77\n",
      "Feature: FLAG_DOCUMENT_9, VIF: 1.21\n",
      "Feature: FLAG_DOCUMENT_10, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_11, VIF: 1.16\n",
      "Feature: FLAG_DOCUMENT_12, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_13, VIF: 1.04\n",
      "Feature: FLAG_DOCUMENT_14, VIF: 1.02\n",
      "Feature: FLAG_DOCUMENT_15, VIF: 1.01\n",
      "Feature: FLAG_DOCUMENT_16, VIF: 1.05\n",
      "Feature: FLAG_DOCUMENT_17, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_18, VIF: 1.04\n",
      "Feature: FLAG_DOCUMENT_19, VIF: 1.00\n",
      "Feature: FLAG_DOCUMENT_20, VIF: 1.02\n",
      "Feature: FLAG_DOCUMENT_21, VIF: 1.01\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_HOUR, VIF: 1.07\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_DAY, VIF: 1.12\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_WEEK, VIF: 1.08\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_MON, VIF: 1.09\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_QRT, VIF: 1.11\n",
      "Feature: AMT_REQ_CREDIT_BUREAU_YEAR, VIF: 1.28\n",
      "Feature: FONDKAPREMONT_MODE_org spec account, VIF: 1.07\n",
      "Feature: FONDKAPREMONT_MODE_reg oper account, VIF: 1.60\n",
      "Feature: FONDKAPREMONT_MODE_reg oper spec account, VIF: 1.10\n",
      "Feature: HOUSETYPE_MODE_ENCODED, VIF: 1.05\n",
      "Feature: WALLSMATERIAL_MODE_ENCODED, VIF: 1.37\n",
      "Feature: NAME_INCOME_TYPE_1.0, VIF: 1.57\n",
      "Feature: NAME_INCOME_TYPE_2.0, VIF: 4.36\n",
      "Feature: NAME_INCOME_TYPE_3.0, VIF: 1.01\n",
      "Feature: NAME_FAMILY_STATUS_2.0, VIF: inf\n",
      "Feature: NAME_FAMILY_STATUS_3.0, VIF: inf\n",
      "\n",
      "VIF 값이 임계값을 초과하는 변수들:\n",
      "                        Feature           VIF\n",
      "4                  CNT_CHILDREN           inf\n",
      "6                    AMT_CREDIT  4.420885e+01\n",
      "8               AMT_GOODS_PRICE  4.581174e+01\n",
      "14                DAYS_EMPLOYED  2.568540e+03\n",
      "18                   FLAG_MOBIL  1.505419e+04\n",
      "19               FLAG_EMP_PHONE  1.187813e+04\n",
      "21             FLAG_CONT_MOBILE  5.695607e+02\n",
      "25              CNT_FAM_MEMBERS           inf\n",
      "26         REGION_RATING_CLIENT  1.101880e+01\n",
      "27  REGION_RATING_CLIENT_W_CITY  1.089233e+01\n",
      "56     OBS_30_CNT_SOCIAL_CIRCLE  4.483153e+02\n",
      "58     OBS_60_CNT_SOCIAL_CIRCLE  4.482906e+02\n",
      "62              FLAG_DOCUMENT_3  4.233480e+01\n",
      "95       NAME_FAMILY_STATUS_2.0           inf\n",
      "96       NAME_FAMILY_STATUS_3.0           inf\n",
      "\n",
      "VIF가 높은 변수들 리스트:\n",
      "['CNT_CHILDREN', 'AMT_CREDIT', 'AMT_GOODS_PRICE', 'DAYS_EMPLOYED', 'FLAG_MOBIL', 'FLAG_EMP_PHONE', 'FLAG_CONT_MOBILE', 'CNT_FAM_MEMBERS', 'REGION_RATING_CLIENT', 'REGION_RATING_CLIENT_W_CITY', 'OBS_30_CNT_SOCIAL_CIRCLE', 'OBS_60_CNT_SOCIAL_CIRCLE', 'FLAG_DOCUMENT_3', 'NAME_FAMILY_STATUS_2.0', 'NAME_FAMILY_STATUS_3.0']\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
    "from statsmodels.tools.tools import add_constant\n",
    "\n",
    "def check_vif_live(df, thresh=10):\n",
    "    \"\"\"\n",
    "    VIF 값을 계산하고 한 번 계산할 때마다 출력하는 함수.\n",
    "    Parameters:\n",
    "        - df (pd.DataFrame): VIF 계산 대상 데이터프레임\n",
    "        - thresh (float): VIF 임계값 (기본값: 8)\n",
    "    Returns:\n",
    "        - pd.DataFrame: VIF가 임계값을 초과하는 변수들 및 VIF 값\n",
    "    \"\"\"\n",
    "    vif_df = pd.DataFrame(columns=[\"Feature\", \"VIF\"])\n",
    "    \n",
    "    print(\"VIF 계산 시작...\\n\")\n",
    "    \n",
    "    # VIF 계산\n",
    "    for i in range(df.shape[1]):\n",
    "        feature_name = df.columns[i]\n",
    "        vif_value = variance_inflation_factor(df.values, i)\n",
    "        vif_df = pd.concat([vif_df, pd.DataFrame({\"Feature\": [feature_name], \"VIF\": [vif_value]})], ignore_index=True)\n",
    "        \n",
    "        # 계산한 VIF 바로 출력\n",
    "        print(f\"Feature: {feature_name}, VIF: {vif_value:.2f}\")\n",
    "    \n",
    "    # VIF가 임계값을 초과하는 변수 필터링\n",
    "    high_vif = vif_df[vif_df[\"VIF\"] > thresh]\n",
    "    \n",
    "    print(\"\\nVIF 값이 임계값을 초과하는 변수들:\")\n",
    "    print(high_vif)\n",
    "\n",
    "    return high_vif\n",
    "\n",
    "# 예제: scaled_df에 VIF 확인\n",
    "high_vif_features = check_vif_live(df_drop_avg_medi_data)\n",
    "\n",
    "# VIF가 높은 변수 리스트 출력\n",
    "print(\"\\nVIF가 높은 변수들 리스트:\")\n",
    "print(high_vif_features[\"Feature\"].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "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>Feature</th>\n",
       "      <th>VIF</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CNT_CHILDREN</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>AMT_CREDIT</td>\n",
       "      <td>4.420885e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>AMT_GOODS_PRICE</td>\n",
       "      <td>4.581174e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>DAYS_EMPLOYED</td>\n",
       "      <td>2.568540e+03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>FLAG_MOBIL</td>\n",
       "      <td>1.505419e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>FLAG_EMP_PHONE</td>\n",
       "      <td>1.187813e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>FLAG_CONT_MOBILE</td>\n",
       "      <td>5.695607e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>CNT_FAM_MEMBERS</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>REGION_RATING_CLIENT</td>\n",
       "      <td>1.101880e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>REGION_RATING_CLIENT_W_CITY</td>\n",
       "      <td>1.089233e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>OBS_30_CNT_SOCIAL_CIRCLE</td>\n",
       "      <td>4.483153e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>OBS_60_CNT_SOCIAL_CIRCLE</td>\n",
       "      <td>4.482906e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>FLAG_DOCUMENT_3</td>\n",
       "      <td>4.233480e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>NAME_FAMILY_STATUS_2.0</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>NAME_FAMILY_STATUS_3.0</td>\n",
       "      <td>inf</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        Feature           VIF\n",
       "4                  CNT_CHILDREN           inf\n",
       "6                    AMT_CREDIT  4.420885e+01\n",
       "8               AMT_GOODS_PRICE  4.581174e+01\n",
       "14                DAYS_EMPLOYED  2.568540e+03\n",
       "18                   FLAG_MOBIL  1.505419e+04\n",
       "19               FLAG_EMP_PHONE  1.187813e+04\n",
       "21             FLAG_CONT_MOBILE  5.695607e+02\n",
       "25              CNT_FAM_MEMBERS           inf\n",
       "26         REGION_RATING_CLIENT  1.101880e+01\n",
       "27  REGION_RATING_CLIENT_W_CITY  1.089233e+01\n",
       "56     OBS_30_CNT_SOCIAL_CIRCLE  4.483153e+02\n",
       "58     OBS_60_CNT_SOCIAL_CIRCLE  4.482906e+02\n",
       "62              FLAG_DOCUMENT_3  4.233480e+01\n",
       "95       NAME_FAMILY_STATUS_2.0           inf\n",
       "96       NAME_FAMILY_STATUS_3.0           inf"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "high_vif_features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 결측치를 전부 버려봤습니다"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "결측치 제거 완료!\n",
      "제거 전 데이터 개수: 307511\n",
      "제거 후 데이터 개수: 8602\n"
     ]
    }
   ],
   "source": [
    "# 결측치가 있는 행 제거\n",
    "df_cleaned = df.dropna()\n",
    "\n",
    "# 결과 확인\n",
    "print(\"결측치 제거 완료!\")\n",
    "print(f\"제거 전 데이터 개수: {len(df)}\")\n",
    "print(f\"제거 후 데이터 개수: {len(df_cleaned)}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 로지스틱 무지성 도전"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "scaled_df = pd.read_csv(r\"C:\\Users\\Administrator\\Desktop\\scaled_by_skweness_and_kurtosis.csv\")\n",
    "X = scaled_df.drop(columns='TARGET')\n",
    "y = scaled_df['TARGET']\n",
    "X.isna().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install imblearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Before SMOTE:\n",
      "Class distribution in y_train:\n",
      "[197856  17398]\n",
      "\n",
      "After SMOTE:\n",
      "Class distribution in y_train_resampled:\n",
      "[197856 197856]\n"
     ]
    }
   ],
   "source": [
    "from imblearn.over_sampling import SMOTE\n",
    "from sklearn.model_selection import train_test_split\n",
    "import numpy as np\n",
    "\n",
    "# 1. 데이터 분할\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)\n",
    "\n",
    "# 2. SMOTE 적용\n",
    "smote = SMOTE(random_state=42)\n",
    "X_train_resampled, y_train_resampled = smote.fit_resample(X_train, y_train)\n",
    "\n",
    "# 3. 결과 확인\n",
    "print(\"Before SMOTE:\")\n",
    "print(f\"Class distribution in y_train:\\n{np.bincount(y_train)}\")\n",
    "\n",
    "print(\"\\nAfter SMOTE:\")\n",
    "print(f\"Class distribution in y_train_resampled:\\n{np.bincount(y_train_resampled)}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "import statsmodels.api as sm\n",
    "\n",
    "# statsmodels 학습모델 구축\n",
    "lr = sm.Logit(y_train, X_train)\n",
    "\n",
    "# 모델 적합\n",
    "result = lr.fit()\n",
    "result.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "모델 학습 시작...\n",
      "모델 학습 완료!\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Python312\\Lib\\site-packages\\sklearn\\linear_model\\_sag.py:349: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "#c의 경우 알파값의 역순으로 하여야 함\n",
    "model = LogisticRegression(penalty='l1', C=0.001, solver='saga')  # 또는 solver='saga'\n",
    "print(\"모델 학습 시작...\")\n",
    "model.fit(X_train_resampled, y_train_resampled)\n",
    "print(\"모델 학습 완료!\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "계수가 0인 특성:\n",
      "['FLAG_DOCUMENT_17', 'AMT_REQ_CREDIT_BUREAU_HOUR', 'COMMONAREA_MODE', 'ENTRANCES_MODE', 'LIVINGAPARTMENTS_MODE', 'NONLIVINGAPARTMENTS_MODE', 'YEARS_BUILD_MEDI', 'COMMONAREA_MEDI', 'ENTRANCES_MEDI', 'FLOORSMIN_MEDI', 'NONLIVINGAPARTMENTS_MEDI', 'EMERGENCYSTATE_MODE', 'NAME_INCOME_TYPE_3.0', 'HOUSETYPE_MODE_ENCODED', 'AMT_REQ_CREDIT_BUREAU_WEEK', 'AMT_REQ_CREDIT_BUREAU_DAY', 'FLAG_DOCUMENT_2', 'FLAG_DOCUMENT_16', 'FLAG_DOCUMENT_4', 'FLAG_DOCUMENT_7', 'FLAG_DOCUMENT_9', 'FLAG_DOCUMENT_10', 'FLAG_DOCUMENT_11', 'FLAG_DOCUMENT_12', 'FLAG_DOCUMENT_13', 'FLAG_DOCUMENT_21', 'FLAG_DOCUMENT_20', 'NONLIVINGAPARTMENTS_AVG', 'FLAG_DOCUMENT_19', 'FLAG_DOCUMENT_14', 'FLAG_DOCUMENT_15', 'FLAG_DOCUMENT_18', 'YEARS_BUILD_MODE', 'FLOORSMIN_MODE', 'YEARS_BUILD_AVG', 'COMMONAREA_AVG', 'ENTRANCES_AVG', 'FLOORSMIN_AVG']\n"
     ]
    }
   ],
   "source": [
    "# 각 변수의 계수 출력\n",
    "coef = model.coef_[0]\n",
    "columns = X_train.columns\n",
    "importance_df = pd.DataFrame({'Feature': columns, 'Coefficient': coef})\n",
    "importance_df = importance_df.sort_values(by='Coefficient', ascending=False)\n",
    "\n",
    "# 계수가 0인 항목 필터링\n",
    "zero_coefficients = importance_df[importance_df['Coefficient'] == 0]\n",
    "\n",
    "# 해당 특성 이름을 리스트로 추출\n",
    "zero_features = zero_coefficients['Feature'].tolist()\n",
    "\n",
    "# 결과 출력\n",
    "print(\"계수가 0인 특성:\")\n",
    "print(zero_features)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Evaluation Metrics (Class 1 - Unpaid):\n",
      "Recall (1): 0.737\n",
      "Precision (1): 0.092\n",
      "F1 Score (1): 0.164\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import confusion_matrix, recall_score, precision_score, f1_score\n",
    "\n",
    "# 1. 예측\n",
    "y_pred = model.predict(X_test)  # X_test로 예측값 생성\n",
    "\n",
    "# 2. 혼동 행렬 계산\n",
    "conf_matrix = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "# 3. 혼동 행렬 시각화\n",
    "plt.figure(figsize=(6, 5))\n",
    "sns.heatmap(conf_matrix, annot=True, fmt=\"d\", cmap=\"Blues\", cbar=False, \n",
    "            xticklabels=['Paid (0)', 'Unpaid (1)'],  # 클래스 0, 1\n",
    "            yticklabels=['Paid (0)', 'Unpaid (1)'])\n",
    "\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Actual')\n",
    "plt.show()\n",
    "\n",
    "# 4. Recall, Precision, F1 Score: 클래스 1(안 갚은 사람) 기준\n",
    "recall_1 = recall_score(y_test, y_pred, pos_label=1)  # 클래스 1 기준\n",
    "precision_1 = precision_score(y_test, y_pred, pos_label=1)\n",
    "f1_1 = f1_score(y_test, y_pred, pos_label=1)\n",
    "\n",
    "# 5. 결과 출력\n",
    "print(\"\\nEvaluation Metrics (Class 1 - Unpaid):\")\n",
    "print(f\"Recall (1): {recall_1:.3f}\")      # 실제로 안 갚은 사람 중 예측된 비율\n",
    "print(f\"Precision (1): {precision_1:.3f}\")  # 안 갚았다고 예측된 사람 중 실제 비율\n",
    "print(f\"F1 Score (1): {f1_1:.3f}\")        # Recall과 Precision의 조화 평균\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}