07 multiple_regression.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import yfinance as yf\n",
    "# 한글폰트 출력 가능 코드\n",
    "from matplotlib import rc\n",
    "rc('font', family='AppleGothic') # Mac Os\n",
    "#rc('font', family='NanumGothic') # Windows Os\n",
    "plt.rcParams['axes.unicode_minus'] = False"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 데이터 불러오기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[*********************100%***********************]  5 of 5 completed\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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"5\" halign=\"left\">Adj Close</th>\n",
       "      <th colspan=\"5\" halign=\"left\">Close</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"5\" halign=\"left\">Open</th>\n",
       "      <th colspan=\"5\" halign=\"left\">Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>000660.KS</th>\n",
       "      <th>005930.KS</th>\n",
       "      <th>035420.KS</th>\n",
       "      <th>035720.KS</th>\n",
       "      <th>^KS11</th>\n",
       "      <th>000660.KS</th>\n",
       "      <th>005930.KS</th>\n",
       "      <th>035420.KS</th>\n",
       "      <th>035720.KS</th>\n",
       "      <th>^KS11</th>\n",
       "      <th>...</th>\n",
       "      <th>000660.KS</th>\n",
       "      <th>005930.KS</th>\n",
       "      <th>035420.KS</th>\n",
       "      <th>035720.KS</th>\n",
       "      <th>^KS11</th>\n",
       "      <th>000660.KS</th>\n",
       "      <th>005930.KS</th>\n",
       "      <th>035420.KS</th>\n",
       "      <th>035720.KS</th>\n",
       "      <th>^KS11</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2021-12-01</th>\n",
       "      <td>115092.867188</td>\n",
       "      <td>74065.523438</td>\n",
       "      <td>389481.68750</td>\n",
       "      <td>122442.796875</td>\n",
       "      <td>2899.719971</td>\n",
       "      <td>116500.0</td>\n",
       "      <td>74400.0</td>\n",
       "      <td>390000.0</td>\n",
       "      <td>122500.0</td>\n",
       "      <td>2899.719971</td>\n",
       "      <td>...</td>\n",
       "      <td>114500.0</td>\n",
       "      <td>72000.0</td>\n",
       "      <td>385000.0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>2860.120117</td>\n",
       "      <td>4214135</td>\n",
       "      <td>21954856</td>\n",
       "      <td>548840</td>\n",
       "      <td>1482581</td>\n",
       "      <td>563100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-02</th>\n",
       "      <td>118550.593750</td>\n",
       "      <td>75459.234375</td>\n",
       "      <td>397970.40625</td>\n",
       "      <td>122442.796875</td>\n",
       "      <td>2945.270020</td>\n",
       "      <td>120000.0</td>\n",
       "      <td>75800.0</td>\n",
       "      <td>398500.0</td>\n",
       "      <td>122500.0</td>\n",
       "      <td>2945.270020</td>\n",
       "      <td>...</td>\n",
       "      <td>115500.0</td>\n",
       "      <td>73900.0</td>\n",
       "      <td>382500.0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>2874.639893</td>\n",
       "      <td>6980518</td>\n",
       "      <td>23652940</td>\n",
       "      <td>613006</td>\n",
       "      <td>2028101</td>\n",
       "      <td>533700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-03</th>\n",
       "      <td>116574.750000</td>\n",
       "      <td>75260.132812</td>\n",
       "      <td>401465.75000</td>\n",
       "      <td>123442.335938</td>\n",
       "      <td>2968.330078</td>\n",
       "      <td>118000.0</td>\n",
       "      <td>75600.0</td>\n",
       "      <td>402000.0</td>\n",
       "      <td>123500.0</td>\n",
       "      <td>2968.330078</td>\n",
       "      <td>...</td>\n",
       "      <td>119000.0</td>\n",
       "      <td>75600.0</td>\n",
       "      <td>397500.0</td>\n",
       "      <td>122000.0</td>\n",
       "      <td>2935.929932</td>\n",
       "      <td>4567843</td>\n",
       "      <td>18330240</td>\n",
       "      <td>501099</td>\n",
       "      <td>1488902</td>\n",
       "      <td>486100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-06</th>\n",
       "      <td>117068.710938</td>\n",
       "      <td>75956.984375</td>\n",
       "      <td>391479.03125</td>\n",
       "      <td>120943.500000</td>\n",
       "      <td>2973.250000</td>\n",
       "      <td>118500.0</td>\n",
       "      <td>76300.0</td>\n",
       "      <td>392000.0</td>\n",
       "      <td>121000.0</td>\n",
       "      <td>2973.250000</td>\n",
       "      <td>...</td>\n",
       "      <td>116000.0</td>\n",
       "      <td>75100.0</td>\n",
       "      <td>400500.0</td>\n",
       "      <td>122500.0</td>\n",
       "      <td>2954.820068</td>\n",
       "      <td>4318893</td>\n",
       "      <td>16391250</td>\n",
       "      <td>456773</td>\n",
       "      <td>1477396</td>\n",
       "      <td>479400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-07</th>\n",
       "      <td>120032.476562</td>\n",
       "      <td>77052.039062</td>\n",
       "      <td>390480.34375</td>\n",
       "      <td>120443.734375</td>\n",
       "      <td>2991.719971</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>77400.0</td>\n",
       "      <td>391000.0</td>\n",
       "      <td>120500.0</td>\n",
       "      <td>2991.719971</td>\n",
       "      <td>...</td>\n",
       "      <td>119000.0</td>\n",
       "      <td>76100.0</td>\n",
       "      <td>390500.0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>2973.840088</td>\n",
       "      <td>5112025</td>\n",
       "      <td>19232453</td>\n",
       "      <td>384366</td>\n",
       "      <td>1218740</td>\n",
       "      <td>540900</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Adj Close                                             \\\n",
       "                000660.KS     005930.KS     035420.KS      035720.KS   \n",
       "Date                                                                   \n",
       "2021-12-01  115092.867188  74065.523438  389481.68750  122442.796875   \n",
       "2021-12-02  118550.593750  75459.234375  397970.40625  122442.796875   \n",
       "2021-12-03  116574.750000  75260.132812  401465.75000  123442.335938   \n",
       "2021-12-06  117068.710938  75956.984375  391479.03125  120943.500000   \n",
       "2021-12-07  120032.476562  77052.039062  390480.34375  120443.734375   \n",
       "\n",
       "                            Close                                             \\\n",
       "                  ^KS11 000660.KS 005930.KS 035420.KS 035720.KS        ^KS11   \n",
       "Date                                                                           \n",
       "2021-12-01  2899.719971  116500.0   74400.0  390000.0  122500.0  2899.719971   \n",
       "2021-12-02  2945.270020  120000.0   75800.0  398500.0  122500.0  2945.270020   \n",
       "2021-12-03  2968.330078  118000.0   75600.0  402000.0  123500.0  2968.330078   \n",
       "2021-12-06  2973.250000  118500.0   76300.0  392000.0  121000.0  2973.250000   \n",
       "2021-12-07  2991.719971  121500.0   77400.0  391000.0  120500.0  2991.719971   \n",
       "\n",
       "            ...      Open                                             \\\n",
       "            ... 000660.KS 005930.KS 035420.KS 035720.KS        ^KS11   \n",
       "Date        ...                                                        \n",
       "2021-12-01  ...  114500.0   72000.0  385000.0  121500.0  2860.120117   \n",
       "2021-12-02  ...  115500.0   73900.0  382500.0  121500.0  2874.639893   \n",
       "2021-12-03  ...  119000.0   75600.0  397500.0  122000.0  2935.929932   \n",
       "2021-12-06  ...  116000.0   75100.0  400500.0  122500.0  2954.820068   \n",
       "2021-12-07  ...  119000.0   76100.0  390500.0  121500.0  2973.840088   \n",
       "\n",
       "              Volume                                        \n",
       "           000660.KS 005930.KS 035420.KS 035720.KS   ^KS11  \n",
       "Date                                                        \n",
       "2021-12-01   4214135  21954856    548840   1482581  563100  \n",
       "2021-12-02   6980518  23652940    613006   2028101  533700  \n",
       "2021-12-03   4567843  18330240    501099   1488902  486100  \n",
       "2021-12-06   4318893  16391250    456773   1477396  479400  \n",
       "2021-12-07   5112025  19232453    384366   1218740  540900  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 삼성전자, 하이닉스, 네이버, 카카오 KOSPI 데이터 불러오기\n",
    "stock_price = yf.download({'005930.KS','000660.KS','035420.KS','035720.KS','^KS11'}, '2021-12-01', '2022-01-21')\n",
    "\n",
    "stock_price.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 35 entries, 2021-12-01 to 2022-01-20\n",
      "Data columns (total 30 columns):\n",
      " #   Column                  Non-Null Count  Dtype  \n",
      "---  ------                  --------------  -----  \n",
      " 0   (Adj Close, 000660.KS)  35 non-null     float64\n",
      " 1   (Adj Close, 005930.KS)  35 non-null     float64\n",
      " 2   (Adj Close, 035420.KS)  35 non-null     float64\n",
      " 3   (Adj Close, 035720.KS)  35 non-null     float64\n",
      " 4   (Adj Close, ^KS11)      35 non-null     float64\n",
      " 5   (Close, 000660.KS)      35 non-null     float64\n",
      " 6   (Close, 005930.KS)      35 non-null     float64\n",
      " 7   (Close, 035420.KS)      35 non-null     float64\n",
      " 8   (Close, 035720.KS)      35 non-null     float64\n",
      " 9   (Close, ^KS11)          35 non-null     float64\n",
      " 10  (High, 000660.KS)       35 non-null     float64\n",
      " 11  (High, 005930.KS)       35 non-null     float64\n",
      " 12  (High, 035420.KS)       35 non-null     float64\n",
      " 13  (High, 035720.KS)       35 non-null     float64\n",
      " 14  (High, ^KS11)           35 non-null     float64\n",
      " 15  (Low, 000660.KS)        35 non-null     float64\n",
      " 16  (Low, 005930.KS)        35 non-null     float64\n",
      " 17  (Low, 035420.KS)        35 non-null     float64\n",
      " 18  (Low, 035720.KS)        35 non-null     float64\n",
      " 19  (Low, ^KS11)            35 non-null     float64\n",
      " 20  (Open, 000660.KS)       35 non-null     float64\n",
      " 21  (Open, 005930.KS)       35 non-null     float64\n",
      " 22  (Open, 035420.KS)       35 non-null     float64\n",
      " 23  (Open, 035720.KS)       35 non-null     float64\n",
      " 24  (Open, ^KS11)           35 non-null     float64\n",
      " 25  (Volume, 000660.KS)     35 non-null     int64  \n",
      " 26  (Volume, 005930.KS)     35 non-null     int64  \n",
      " 27  (Volume, 035420.KS)     35 non-null     int64  \n",
      " 28  (Volume, 035720.KS)     35 non-null     int64  \n",
      " 29  (Volume, ^KS11)         35 non-null     int64  \n",
      "dtypes: float64(25), int64(5)\n",
      "memory usage: 8.5 KB\n"
     ]
    }
   ],
   "source": [
    "stock_price.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "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>Adj Close_000660.KS</th>\n",
       "      <th>Adj Close_005930.KS</th>\n",
       "      <th>Adj Close_035420.KS</th>\n",
       "      <th>Adj Close_035720.KS</th>\n",
       "      <th>Adj Close_^KS11</th>\n",
       "      <th>Close_000660.KS</th>\n",
       "      <th>Close_005930.KS</th>\n",
       "      <th>Close_035420.KS</th>\n",
       "      <th>Close_035720.KS</th>\n",
       "      <th>Close_^KS11</th>\n",
       "      <th>...</th>\n",
       "      <th>Open_000660.KS</th>\n",
       "      <th>Open_005930.KS</th>\n",
       "      <th>Open_035420.KS</th>\n",
       "      <th>Open_035720.KS</th>\n",
       "      <th>Open_^KS11</th>\n",
       "      <th>Volume_000660.KS</th>\n",
       "      <th>Volume_005930.KS</th>\n",
       "      <th>Volume_035420.KS</th>\n",
       "      <th>Volume_035720.KS</th>\n",
       "      <th>Volume_^KS11</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2021-12-01</th>\n",
       "      <td>115092.867188</td>\n",
       "      <td>74065.523438</td>\n",
       "      <td>389481.68750</td>\n",
       "      <td>122442.796875</td>\n",
       "      <td>2899.719971</td>\n",
       "      <td>116500.0</td>\n",
       "      <td>74400.0</td>\n",
       "      <td>390000.0</td>\n",
       "      <td>122500.0</td>\n",
       "      <td>2899.719971</td>\n",
       "      <td>...</td>\n",
       "      <td>114500.0</td>\n",
       "      <td>72000.0</td>\n",
       "      <td>385000.0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>2860.120117</td>\n",
       "      <td>4214135</td>\n",
       "      <td>21954856</td>\n",
       "      <td>548840</td>\n",
       "      <td>1482581</td>\n",
       "      <td>563100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-02</th>\n",
       "      <td>118550.593750</td>\n",
       "      <td>75459.234375</td>\n",
       "      <td>397970.40625</td>\n",
       "      <td>122442.796875</td>\n",
       "      <td>2945.270020</td>\n",
       "      <td>120000.0</td>\n",
       "      <td>75800.0</td>\n",
       "      <td>398500.0</td>\n",
       "      <td>122500.0</td>\n",
       "      <td>2945.270020</td>\n",
       "      <td>...</td>\n",
       "      <td>115500.0</td>\n",
       "      <td>73900.0</td>\n",
       "      <td>382500.0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>2874.639893</td>\n",
       "      <td>6980518</td>\n",
       "      <td>23652940</td>\n",
       "      <td>613006</td>\n",
       "      <td>2028101</td>\n",
       "      <td>533700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-03</th>\n",
       "      <td>116574.750000</td>\n",
       "      <td>75260.132812</td>\n",
       "      <td>401465.75000</td>\n",
       "      <td>123442.335938</td>\n",
       "      <td>2968.330078</td>\n",
       "      <td>118000.0</td>\n",
       "      <td>75600.0</td>\n",
       "      <td>402000.0</td>\n",
       "      <td>123500.0</td>\n",
       "      <td>2968.330078</td>\n",
       "      <td>...</td>\n",
       "      <td>119000.0</td>\n",
       "      <td>75600.0</td>\n",
       "      <td>397500.0</td>\n",
       "      <td>122000.0</td>\n",
       "      <td>2935.929932</td>\n",
       "      <td>4567843</td>\n",
       "      <td>18330240</td>\n",
       "      <td>501099</td>\n",
       "      <td>1488902</td>\n",
       "      <td>486100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-06</th>\n",
       "      <td>117068.710938</td>\n",
       "      <td>75956.984375</td>\n",
       "      <td>391479.03125</td>\n",
       "      <td>120943.500000</td>\n",
       "      <td>2973.250000</td>\n",
       "      <td>118500.0</td>\n",
       "      <td>76300.0</td>\n",
       "      <td>392000.0</td>\n",
       "      <td>121000.0</td>\n",
       "      <td>2973.250000</td>\n",
       "      <td>...</td>\n",
       "      <td>116000.0</td>\n",
       "      <td>75100.0</td>\n",
       "      <td>400500.0</td>\n",
       "      <td>122500.0</td>\n",
       "      <td>2954.820068</td>\n",
       "      <td>4318893</td>\n",
       "      <td>16391250</td>\n",
       "      <td>456773</td>\n",
       "      <td>1477396</td>\n",
       "      <td>479400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-12-07</th>\n",
       "      <td>120032.476562</td>\n",
       "      <td>77052.039062</td>\n",
       "      <td>390480.34375</td>\n",
       "      <td>120443.734375</td>\n",
       "      <td>2991.719971</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>77400.0</td>\n",
       "      <td>391000.0</td>\n",
       "      <td>120500.0</td>\n",
       "      <td>2991.719971</td>\n",
       "      <td>...</td>\n",
       "      <td>119000.0</td>\n",
       "      <td>76100.0</td>\n",
       "      <td>390500.0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>2973.840088</td>\n",
       "      <td>5112025</td>\n",
       "      <td>19232453</td>\n",
       "      <td>384366</td>\n",
       "      <td>1218740</td>\n",
       "      <td>540900</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Adj Close_000660.KS  Adj Close_005930.KS  Adj Close_035420.KS  \\\n",
       "Date                                                                        \n",
       "2021-12-01        115092.867188         74065.523438         389481.68750   \n",
       "2021-12-02        118550.593750         75459.234375         397970.40625   \n",
       "2021-12-03        116574.750000         75260.132812         401465.75000   \n",
       "2021-12-06        117068.710938         75956.984375         391479.03125   \n",
       "2021-12-07        120032.476562         77052.039062         390480.34375   \n",
       "\n",
       "            Adj Close_035720.KS  Adj Close_^KS11  Close_000660.KS  \\\n",
       "Date                                                                \n",
       "2021-12-01        122442.796875      2899.719971         116500.0   \n",
       "2021-12-02        122442.796875      2945.270020         120000.0   \n",
       "2021-12-03        123442.335938      2968.330078         118000.0   \n",
       "2021-12-06        120943.500000      2973.250000         118500.0   \n",
       "2021-12-07        120443.734375      2991.719971         121500.0   \n",
       "\n",
       "            Close_005930.KS  Close_035420.KS  Close_035720.KS  Close_^KS11  \\\n",
       "Date                                                                         \n",
       "2021-12-01          74400.0         390000.0         122500.0  2899.719971   \n",
       "2021-12-02          75800.0         398500.0         122500.0  2945.270020   \n",
       "2021-12-03          75600.0         402000.0         123500.0  2968.330078   \n",
       "2021-12-06          76300.0         392000.0         121000.0  2973.250000   \n",
       "2021-12-07          77400.0         391000.0         120500.0  2991.719971   \n",
       "\n",
       "            ...  Open_000660.KS  Open_005930.KS  Open_035420.KS  \\\n",
       "Date        ...                                                   \n",
       "2021-12-01  ...        114500.0         72000.0        385000.0   \n",
       "2021-12-02  ...        115500.0         73900.0        382500.0   \n",
       "2021-12-03  ...        119000.0         75600.0        397500.0   \n",
       "2021-12-06  ...        116000.0         75100.0        400500.0   \n",
       "2021-12-07  ...        119000.0         76100.0        390500.0   \n",
       "\n",
       "            Open_035720.KS   Open_^KS11  Volume_000660.KS  Volume_005930.KS  \\\n",
       "Date                                                                          \n",
       "2021-12-01        121500.0  2860.120117           4214135          21954856   \n",
       "2021-12-02        121500.0  2874.639893           6980518          23652940   \n",
       "2021-12-03        122000.0  2935.929932           4567843          18330240   \n",
       "2021-12-06        122500.0  2954.820068           4318893          16391250   \n",
       "2021-12-07        121500.0  2973.840088           5112025          19232453   \n",
       "\n",
       "            Volume_035420.KS  Volume_035720.KS  Volume_^KS11  \n",
       "Date                                                          \n",
       "2021-12-01            548840           1482581        563100  \n",
       "2021-12-02            613006           2028101        533700  \n",
       "2021-12-03            501099           1488902        486100  \n",
       "2021-12-06            456773           1477396        479400  \n",
       "2021-12-07            384366           1218740        540900  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 컬럼 명 수정하기\n",
    "col_names = stock_price.columns\n",
    "new_names = []\n",
    "for i in range(len(col_names)):\n",
    "    new_names.append(col_names[i][0] + '_' + col_names[i][1])\n",
    "stock_price.columns = new_names\n",
    "stock_price.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. 다중회귀분석"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 필요한 라이브러리\n",
    "import statsmodels.api as sm\n",
    "\n",
    "# target은 종속변수, feature는 독립변수와 같다. (KOSPI 제외 Close, Volume 컬럼에 대해 feature 설정)\n",
    "target = stock_price['Close_^KS11']\n",
    "feature1 = stock_price[new_names[5:9]]\n",
    "feature2 = stock_price[new_names[-5:-1]]\n",
    "feature = pd.concat([feature1, feature2], axis=1)\n",
    "\n",
    "# 상수항을 더해줘서 회귀식을 더 잘 적합할 수 있도록 만들어 준다.\n",
    "feature_add = sm.add_constant(feature, has_constant='add')\n",
    "# 상수항 더한 것 확인\n",
    "# feature_add"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>       <td>Close_^KS11</td>   <th>  R-squared:         </th> <td>   0.883</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.846</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   24.43</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Wed, 30 Mar 2022</td> <th>  Prob (F-statistic):</th> <td>3.19e-10</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>16:49:45</td>     <th>  Log-Likelihood:    </th> <td> -147.05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td>    35</td>      <th>  AIC:               </th> <td>   312.1</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td>    26</td>      <th>  BIC:               </th> <td>   326.1</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>     8</td>      <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "          <td></td>            <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th>            <td>  808.6964</td> <td>  258.310</td> <td>    3.131</td> <td> 0.004</td> <td>  277.732</td> <td> 1339.660</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Close_000660.KS</th>  <td>    0.0010</td> <td>    0.002</td> <td>    0.487</td> <td> 0.630</td> <td>   -0.003</td> <td>    0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Close_005930.KS</th>  <td>    0.0184</td> <td>    0.005</td> <td>    4.021</td> <td> 0.000</td> <td>    0.009</td> <td>    0.028</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Close_035420.KS</th>  <td>    0.0008</td> <td>    0.001</td> <td>    0.855</td> <td> 0.400</td> <td>   -0.001</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Close_035720.KS</th>  <td>    0.0028</td> <td>    0.002</td> <td>    1.563</td> <td> 0.130</td> <td>   -0.001</td> <td>    0.007</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Volume_000660.KS</th> <td>-3.328e-07</td> <td> 5.14e-06</td> <td>   -0.065</td> <td> 0.949</td> <td>-1.09e-05</td> <td> 1.02e-05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Volume_005930.KS</th> <td>-1.633e-06</td> <td> 1.43e-06</td> <td>   -1.139</td> <td> 0.265</td> <td>-4.58e-06</td> <td> 1.31e-06</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Volume_035420.KS</th> <td> 5.354e-05</td> <td> 3.07e-05</td> <td>    1.742</td> <td> 0.093</td> <td>-9.62e-06</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Volume_035720.KS</th> <td>-8.976e-07</td> <td> 5.81e-06</td> <td>   -0.155</td> <td> 0.878</td> <td>-1.28e-05</td> <td>  1.1e-05</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td> 0.662</td> <th>  Durbin-Watson:     </th> <td>   0.916</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th> <td> 0.718</td> <th>  Jarque-Bera (JB):  </th> <td>   0.695</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>          <td> 0.293</td> <th>  Prob(JB):          </th> <td>   0.706</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>      <td> 2.634</td> <th>  Cond. No.          </th> <td>1.29e+09</td>\n",
       "</tr>\n",
       "</table><br/><br/>Notes:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 1.29e+09. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:            Close_^KS11   R-squared:                       0.883\n",
       "Model:                            OLS   Adj. R-squared:                  0.846\n",
       "Method:                 Least Squares   F-statistic:                     24.43\n",
       "Date:                Wed, 30 Mar 2022   Prob (F-statistic):           3.19e-10\n",
       "Time:                        16:49:45   Log-Likelihood:                -147.05\n",
       "No. Observations:                  35   AIC:                             312.1\n",
       "Df Residuals:                      26   BIC:                             326.1\n",
       "Df Model:                           8                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "====================================================================================\n",
       "                       coef    std err          t      P>|t|      [0.025      0.975]\n",
       "------------------------------------------------------------------------------------\n",
       "const              808.6964    258.310      3.131      0.004     277.732    1339.660\n",
       "Close_000660.KS      0.0010      0.002      0.487      0.630      -0.003       0.005\n",
       "Close_005930.KS      0.0184      0.005      4.021      0.000       0.009       0.028\n",
       "Close_035420.KS      0.0008      0.001      0.855      0.400      -0.001       0.003\n",
       "Close_035720.KS      0.0028      0.002      1.563      0.130      -0.001       0.007\n",
       "Volume_000660.KS -3.328e-07   5.14e-06     -0.065      0.949   -1.09e-05    1.02e-05\n",
       "Volume_005930.KS -1.633e-06   1.43e-06     -1.139      0.265   -4.58e-06    1.31e-06\n",
       "Volume_035420.KS  5.354e-05   3.07e-05      1.742      0.093   -9.62e-06       0.000\n",
       "Volume_035720.KS -8.976e-07   5.81e-06     -0.155      0.878   -1.28e-05     1.1e-05\n",
       "==============================================================================\n",
       "Omnibus:                        0.662   Durbin-Watson:                   0.916\n",
       "Prob(Omnibus):                  0.718   Jarque-Bera (JB):                0.695\n",
       "Skew:                           0.293   Prob(JB):                        0.706\n",
       "Kurtosis:                       2.634   Cond. No.                     1.29e+09\n",
       "==============================================================================\n",
       "\n",
       "Notes:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "[2] The condition number is large, 1.29e+09. This might indicate that there are\n",
       "strong multicollinearity or other numerical problems.\n",
       "\"\"\""
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# sm OLS 적합\n",
    "model = sm.OLS(target, feature_add)\n",
    "fitted_model = model.fit()\n",
    "\n",
    "# summary 함수통해 결과출력\n",
    "fitted_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1-1 다중공선성 측정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "feature Close_000660.KS의 VIF는 5.923930810480606입니다.\n"
     ]
    }
   ],
   "source": [
    "# feature 전체에 대해 VIF를 알고 싶은 특정 feature를 target으로 두고 나머지를 feature로 설정하여 OLS를 구한 뒤, 결정계수를 도출한다.\n",
    "calculate_vif = sm.OLS(feature_add['Close_000660.KS'], feature_add.drop(columns = 'Close_000660.KS')).fit()\n",
    "\n",
    "# 다중공선성 계산 : 1 / (1 - R^2)\n",
    "print(f'feature Close_000660.KS의 VIF는 {1 / (1 - calculate_vif.rsquared)}입니다.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "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>VIF Factor</th>\n",
       "      <th>features</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6643.829566</td>\n",
       "      <td>const</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>42.701709</td>\n",
       "      <td>Close_035420.KS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38.846064</td>\n",
       "      <td>Close_035720.KS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>6.994746</td>\n",
       "      <td>Volume_035720.KS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.923931</td>\n",
       "      <td>Close_000660.KS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5.863995</td>\n",
       "      <td>Volume_035420.KS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>4.153377</td>\n",
       "      <td>Volume_005930.KS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>3.791508</td>\n",
       "      <td>Volume_000660.KS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>3.747452</td>\n",
       "      <td>Close_005930.KS</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    VIF Factor          features\n",
       "0  6643.829566             const\n",
       "1    42.701709   Close_035420.KS\n",
       "2    38.846064   Close_035720.KS\n",
       "3     6.994746  Volume_035720.KS\n",
       "4     5.923931   Close_000660.KS\n",
       "5     5.863995  Volume_035420.KS\n",
       "6     4.153377  Volume_005930.KS\n",
       "7     3.791508  Volume_000660.KS\n",
       "8     3.747452   Close_005930.KS"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def vif(data):\n",
    "    import pandas as pd\n",
    "    from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
    "\n",
    "    # VIF 출력을 위한 데이터 프레임 형성\n",
    "    vif = pd.DataFrame()\n",
    "\n",
    "    # VIF 값과 각 Feature 이름에 대해 설정\n",
    "    vif[\"VIF Factor\"] = [variance_inflation_factor(data.values, i) for i in range(len(data.columns))]\n",
    "    vif[\"features\"] = data.columns \n",
    "\n",
    "    # VIF 값이 높은 순으로 정렬\n",
    "    vif = vif.sort_values(by=\"VIF Factor\", ascending=False)\n",
    "    vif = vif.reset_index().drop(columns='index')\n",
    "    \n",
    "    return vif\n",
    "\n",
    "vif(feature_add)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1번째 VIF 측정\n",
      "Max VIF feature & value : const,  6643.8296\n",
      "2번째 VIF 측정\n",
      "Max VIF feature & value : Close_000660.KS,  11214.0518\n",
      "3번째 VIF 측정\n",
      "Max VIF feature & value : Close_000660.KS,  7819.5666\n",
      "4번째 VIF 측정\n",
      "Max VIF feature & value : Close_000660.KS,  182.7620\n",
      "5번째 VIF 측정\n",
      "Max VIF feature & value : Close_035420.KS,  31.6371\n",
      "6번째 VIF 측정\n",
      "Max VIF feature & value : Close_000660.KS,  28.3873\n",
      "7번째 VIF 측정\n",
      "Max VIF feature & value : const,  13.7264\n",
      "8번째 VIF 측정\n",
      "Max VIF feature & value : const,  2.9176\n",
      "\n",
      "\n",
      "Volume_005930.KS의 vif는  2.92입니다.\n",
      "Volume_035720.KS의 vif는  2.92입니다.\n"
     ]
    }
   ],
   "source": [
    "# 다중공선성 측정\n",
    "def vif_drop(x, cut_off):\n",
    "    import pandas as pd\n",
    "    from statsmodels.stats.outliers_influence import variance_inflation_factor\n",
    "    # vif 10 초과시 drop을 위한 임계값 설정\n",
    "    thresh = cut_off\n",
    "    # Filter method로 feature selection 진행 후 최종 도출 될 데이터 프레임 형성\n",
    "    output = pd.DataFrame()\n",
    "    # 데이터의 컬럼 개수 설정\n",
    "    k = x.shape[1]\n",
    "    # VIF 측정\n",
    "    vif = [variance_inflation_factor(x.values, i) for i in range(x.shape[1])]\n",
    "    for i in range(1,k):\n",
    "        print(f'{i}번째 VIF 측정')\n",
    "        # VIF 최대 값 선정\n",
    "        a = np.argmax(vif)\n",
    "        print(f'Max VIF feature & value : {x.columns[a]}, {vif[a] : .4f}')\n",
    "        # VIF 최대 값이 임계치를 넘지 않는 경우 break\n",
    "        if (vif[a] <= thresh):\n",
    "            print('\\n')\n",
    "            for q in range(output.shape[1]):\n",
    "                print(f'{output.columns[q]}의 vif는 {vif[q] : .2f}입니다.')\n",
    "            break\n",
    "        # VIF 최대 값이 임계치를 넘는 경우, + 1번째 시도인 경우 : if 문으로 해당 feature 제거 후 다시 vif 측정\n",
    "        if (i == 1):\n",
    "            output = x.drop(x.columns[a], axis = 1)\n",
    "            vif = [variance_inflation_factor(output.values, j) for j in range(output.shape[1])]\n",
    "        # VIF 최대 값이 임계치를 넘는 경우, + 1번째 이후 시도인 경우 : if 문으로 해당 feature 제거 후 다시 vif 측정\n",
    "        elif (i > 1):\n",
    "            output = output.drop(output.columns[a], axis = 1)\n",
    "            vif = [variance_inflation_factor(output.values, j) for j in range(output.shape[1])]\n",
    "    return(output)\n",
    "\n",
    "feature_drop = vif_drop(feature_add, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>       <td>Close_^KS11</td>   <th>  R-squared:         </th> <td>   0.297</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.253</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   6.749</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Wed, 30 Mar 2022</td> <th>  Prob (F-statistic):</th>  <td>0.00359</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>16:49:47</td>     <th>  Log-Likelihood:    </th> <td> -178.38</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td>    35</td>      <th>  AIC:               </th> <td>   362.8</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td>    32</td>      <th>  BIC:               </th> <td>   367.4</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>     2</td>      <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "          <td></td>            <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th>            <td> 2964.8053</td> <td>   26.551</td> <td>  111.663</td> <td> 0.000</td> <td> 2910.722</td> <td> 3018.888</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Volume_005930.KS</th> <td> 2.513e-06</td> <td> 1.55e-06</td> <td>    1.618</td> <td> 0.115</td> <td> -6.5e-07</td> <td> 5.68e-06</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Volume_035720.KS</th> <td>-1.571e-05</td> <td> 4.85e-06</td> <td>   -3.240</td> <td> 0.003</td> <td>-2.56e-05</td> <td>-5.83e-06</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td> 9.119</td> <th>  Durbin-Watson:     </th> <td>   0.360</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th> <td> 0.010</td> <th>  Jarque-Bera (JB):  </th> <td>   8.351</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>          <td>-1.174</td> <th>  Prob(JB):          </th> <td>  0.0154</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>      <td> 3.464</td> <th>  Cond. No.          </th> <td>5.82e+07</td>\n",
       "</tr>\n",
       "</table><br/><br/>Notes:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 5.82e+07. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:            Close_^KS11   R-squared:                       0.297\n",
       "Model:                            OLS   Adj. R-squared:                  0.253\n",
       "Method:                 Least Squares   F-statistic:                     6.749\n",
       "Date:                Wed, 30 Mar 2022   Prob (F-statistic):            0.00359\n",
       "Time:                        16:49:47   Log-Likelihood:                -178.38\n",
       "No. Observations:                  35   AIC:                             362.8\n",
       "Df Residuals:                      32   BIC:                             367.4\n",
       "Df Model:                           2                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "====================================================================================\n",
       "                       coef    std err          t      P>|t|      [0.025      0.975]\n",
       "------------------------------------------------------------------------------------\n",
       "const             2964.8053     26.551    111.663      0.000    2910.722    3018.888\n",
       "Volume_005930.KS  2.513e-06   1.55e-06      1.618      0.115    -6.5e-07    5.68e-06\n",
       "Volume_035720.KS -1.571e-05   4.85e-06     -3.240      0.003   -2.56e-05   -5.83e-06\n",
       "==============================================================================\n",
       "Omnibus:                        9.119   Durbin-Watson:                   0.360\n",
       "Prob(Omnibus):                  0.010   Jarque-Bera (JB):                8.351\n",
       "Skew:                          -1.174   Prob(JB):                       0.0154\n",
       "Kurtosis:                       3.464   Cond. No.                     5.82e+07\n",
       "==============================================================================\n",
       "\n",
       "Notes:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "[2] The condition number is large, 5.82e+07. This might indicate that there are\n",
       "strong multicollinearity or other numerical problems.\n",
       "\"\"\""
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import statsmodels.api as sm\n",
    "\n",
    "# 다중공선성 제거후 모델 형성\n",
    "\n",
    "feature = feature_drop\n",
    "\n",
    "# sm OLS 적합\n",
    "feature_add = sm.add_constant(feature, has_constant='add')\n",
    "\n",
    "model = sm.OLS(target, feature_add)\n",
    "fitted_model = model.fit()\n",
    "\n",
    "# summary 함수통해 결과출력\n",
    "fitted_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. 회귀 식의 적합성 검정 : 잔차분석\n",
    "\n",
    "### 정규성, 자기상관성, 등분산성 검정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAF1CAYAAADvHYE/AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAnK0lEQVR4nO3df9hldVnv8fcHmDEVgQGHTG0YcRxEpVM2yY/E0EpBA7WLk7+uroJsIJMQujA7VoCmghKE5q8xrJN4EIyMPOZgCTOiCDikiJeF0ciY9MMRnJSTP6Lu88daT7Nnz372/q5nr73Xdz3fz+u61jXPXvve9/que+11z95rr722IgIzMyvHPl0PwMzM5suN38ysMG78ZmaFceM3MyuMG7+ZWWHc+M3MCrNf1wOY5JGPfGSsXbu262GYmfXK7bff/vWIWD3qvuwb/9q1a9m2bVvXwzAz6xVJOxa7z4d6zMwK48ZvZlYYN34zs8K48ZuZFcaN38ysMG78ZmaFceM3MyuMG7+ZWWGy/wJXada+5iN73L7noud1NBKbh+HtDd7mNnt+xW9mVhg3fjOzwrjxm5kVxo3fzKwwbvxmZoWZaeOXdLykewduHyFpq6TbJF0raf9ZLt/MzPY2s8Yv6UDgLOCfB2ZfDpwWEU8D/gQ4f1bLNzOz0Wb5iv8twGuBBwEkHQV8JSK2A0TEdcBxM1y+mZmNMJMvcEl6KXB7RPy9pIXZ64C7hkLvl7QqIr4xi3GYzYK/dGV913rjl7QGOCkifn74LiBGPGSveZI2AhsB1qxZ0/YQrWNunGbdmsWhnhcA6yVtkbQFeFL97w5g/VDsIRGxazhBRGyKiA0RsWH16pG/FWxmZkvU+iv+iHgr8NaF25JuiYgT6r/fIumwiNgh6UTgpraXb2Zm4837Im1nA1eqOvB/L3D6nJe/LPhQiZlNY+aNPyKOGfj7TuD4WS/TzMwW52/umpkVxtfjN7Ml82HHfvIrfjOzwrjxm5kVxod6zGzZ8SGo8fyK38ysMG78ZmaF8aEes8IMHwbxIZDyuPFba3xc1awffKjHzKwwbvxmZoVx4zczK4wbv5lZYdz4zcwK48ZvZlYYn85pNgM+tdVy5lf8ZmaFceM3MytM64d6JB0H/CawP/AQ4MKIuF7SEcAm4KHAPwK/EBEPtL18a8aHJMzKM4tj/PsCL42Ib0laBWwFrgcuB06LiO2Sng+cD5w3g+WbmdkYrTf+iLhp4OYu4NuSjgK+EhHb65jrJL267WXbnvxq3vrEz9f5mdkxfkn7AJcAVwDrgLuGQu6v3xGYmdkczaTxSzoUeD/wiYjYBAiIEaGj5iFpo6Rtkrbt3LlzFkM0MyvWLD7cPRx4G3BGRHy1nr0dePZQ6CERsWtUjvo/i00AGzZsGPmfQ9/4bayZ5WIWr/gvAE4faPpExOeA9ZIOA5B0InDTyEebmdlMzeKsnuOAqyUNzns5cDZwpao77gVOn8GybUb8jsVs+ZjFWT3rxtx9fNvLs+Vr3j8R6P/crBT+5q6ZWWHc+M3MCuOrc5rZTPkQWn78it/MrDBu/GZmhXHjNzMrjBu/mVlh3PjNzArjxm9mVhg3fjOzwrjxm5kVxo3fzKww/uauWQ/M+4J1trz5Fb+ZWWH8it/M9uLr6yxvfsVvZlYYN34zs8K48ZuZFcaN38ysMHP/cFfS9wN/DBwMfBP4hYj4p3mPw8ys1NNku3jF/3vABRFxNPCbwCUdjMHMrFiKiPktTDoIuCoiThqY9xHgZRGxa9RjNmzYENu2bdtz5gknzGyMs3LL9vv2mnfM4YdMjEuJmSYu11yj4kbFtKmUWvT5OZYq1+fYPGnr1tsjYsPI++bc+H8EOD0izhqY9zbgvRHx2YF5G4GNAGvWrPnRHTt27PGW7AP/5zVAHk/ENp+s1px38N1KeC52sY65/ic4KSanxv9U4Bcj4tcG5r0NuCIiPjfqMQuv+FO/UJIS10Uum42lbsvluI1KeC52sY597T2SFm388/5w9x5g3dC8w+v5ZmZjLbf/yLoy18YfEfdL+ndJT42Iv5H0Q8B9ix3f7wM/Ec2sb7q4Vs+rgCsk7Q/8G3BaB2MwMyvW3Bt/RHwVeM68l2vLk99x7eZaWCp/c9fMrDBu/GZmhXHjNzMrjH+IxYrg499mu/kVv5lZYdz4zcwK48ZvZlYYH+Mfw8eFzWw58it+M7PCuPGbmRXGjd/MrDBu/GZmhXHjNzMrjBu/mVlh3PjNzArjxm9mVhg3fjOzwrjxm5kVptXGL+k4SR+WdKOkmyU9Z+C+IyRtlXSbpGvr39w1M7M5a/taPfsCL42Ib0laBWwFrq/vuxw4LSK2S3o+cD5wXsvLNzOzCVpt/BFx08DNXcC3JQl4CvCViNhex10n6dVtLrsJX3zNzEo2k2P8kvYBLgGuiIgA1gF3DYXdX78rGPX4jZK2Sdq2c+fOWQzRzKxYrTd+SYcC7wc+ERGbFmYDMSJ81DwiYlNEbIiIDatXr257iGZmRVvyoR5J5wKn1De3RsT5kg4H3gacERFfHQjfDjx7KMUhEbFrqcs3M7OlWXLjj4hLgUuHZl8AnB4R/zoU+zlJ6yUdFhE7JJ0I3ISZmc1d22f1HAdcXX2e+99eHhF3A2cDV9Yf9t4LnN7yss3MLEHbZ/WsG3PfncDxbS7PzMya8zd3zcwK48ZvZlaYto/xm5ktK8vxC59+xW9mVhg3fjOzwrjxm5kVxo3fzKwwbvxmZoVx4zczK4wbv5lZYdz4zcwK48ZvZlYYN34zs8K48ZuZFcaN38ysMG78ZmaFceM3MyuMG7+ZWWFm0vglHS/p3qF5R0jaKuk2SddK2n8WyzYzs/Fab/ySDgTOAv556K7LgdMi4mnAnwDnt71sMzObbBav+N8CvBZ4cGGGpKOAr0TEdoCIuA44bgbLNjOzCVpt/JJeCtweEX8/dNc64K6hefdLWrVIno2StknatnPnzjaHaGZWvNYav6Q1wEkR8e5RdwMxYv6oeUTEpojYEBEbVq9e3dYQzcyMKX5sXdK5wCn1za3AfcB6SVvqeU+q//51YDvw7KEUh0TErqUu38zMlmbJjT8iLgUuHZr91oU/JN0SEScM3F4v6bCI2CHpROCmpS7bzMyWbsmNfwnOBq6UJOBe4PQ5LtvMzGoza/wRcczQ7TuB42e1PDMzS+Nv7pqZFcaN38ysMG78ZmaFmeeHu3Nxz0XP63oIZmZZ8yt+M7PCuPGbmRXGjd/MrDBu/GZmhXHjNzMrjBu/mVlh3PjNzArjxm9mVhg3fjOzwrjxm5kVxo3fzKwwbvxmZoVZdhdpMzPrQp8uEOlX/GZmhWm98UvaR9JFkj4paauk3xi471hJt0r6jKT3SPI7DjOzOZvFK/7XAN+IiKdHxE8A1wBIWgFcDDw3In4M+AJwxgyWb2ZmY7Ta+CWtBH4WePPCvIj4cv3nicDmiLivvv1O4NQ2l29mZpO1/Yr/ccDngVdKuqGefqq+bx1w10JgRHwPWDEqiaSNkrZJ2rZz586Wh2hmVra2G//+wE8D/xIRzwJeDLxF0ipAQAzFD9+uZkZsiogNEbFh9erVLQ/RzKxsU324Kulc4JT65lbgMuBvI+KDABHxNUmfAh4PbAeeOPDYlcCD0yzfzMyam+oVf0RcGhEn1NP5EbEL+BdJJwJI2h94KtUhns3AyfWrf4DTqT/4NTOz+ZnF6ZS/BrxN0mvq278VEd8CkPRaYLOkB4E7gLNmsHwzMxuj9cZfv+r/+UXu2wIc3fYyzcwsnb+5a2ZWGDd+M7PCuPGbmRXGjd/MrDBu/GZmhXHjNzMrjBu/mVlh3PjNzArjxm9mVhg3fjOzwrjxm5kVxo3fzKwwbvxmZoVx4zczK4wbv5lZYdz4zcwK48ZvZlaYWfz0opmZTeGei5430/ytv+KXtEbSRyRtkXSbpBcO3HespFslfUbSeyT5Px4zszmbxaGeNwNviIgTgJOAi1VZAVwMPDcifgz4AnDGDJZvZmZjzOIV977AV+u/HwC2R0RIOhHYHBH31fe9E7geePsMxmBmtqxNczhoFo3/1cBlkj4JPAs4t56/DrhrISgivle/C9iLpI3ARoA1a9bMYIhmZuWaxaGelwGbqV7RXwi8SdJjAAExFDt8u5oZsSkiNkTEhtWrV89giGZm5ZrqFb+kc4FT6ptbgQ8Ah0fE79bztkn6AHAqsB144sBjVwIPTrN8MzNrbqpX/BFxaUScUE/nA98EjpL0CABJDwF+Dvgc1buAkyWtqh9+OnDNNMs3M7PmWj3GHxH3SnojsFnS96j+Y/nDiNgKIOm19X0PAncAZ7W5fDMzm6z1D3cj4kPAhxa5bwtwdNvLNDOzdL5kg5lZYdz4zcwK48ZvZlYYN34zs8K48ZuZFcaN38ysMG78ZmaFceM3MyuMG7+ZWWHc+M3MCuPGb2ZWGDd+M7PCuPGbmRXGjd/MrDBu/GZmhXHjNzMrjBu/mVlhWv8Frlm556LndT0EM7NlYapX/JJ+U9KZQ/OOkLRV0m2SrpW0/8B9x0q6VdJnJL1HUm/+4zEzWy6W1HglPQa4Fng88NtDd18OnBYR2yU9HzgfOE/SCuBi4LkRcZ+ks4EzgLcvefRmZj2Sy5GLJb3ij4h7I+IY4LzB+ZKOAr4SEdvruOuA4+q7TwQ2R8R99e13AqcuadRmZrZkbX+4uw64a2je/ZJWDd8XEd8DVoxKImmjpG2Stu3cubPlIZqZla3txi8gRsyPRe4bFUtEbIqIDRGxYfXq1S0P0cysbBMbv6RzJW2ppwsnhG8H1g/NOyQidg3fJ2kl8GDD8ZqZ2ZQmNv6IuDQiTqin8yfEfg5YL+kwAEknAjfVd28GTq4P+wCcDlyz5JGbmdmSzOJ0yrOBKyUJuJeqwRMR35H0WmCzpAeBO4CzZrB8MzMbY6rGHxF/PGLencDxi8RvAY6eZplmZjYdX7LBzKwwbvxmZoVRxMgzKrMhaSewY2j2I4GvT3hoSkzfc3WxzFxzdbHMEnJ1scxcc3WxzGlyHRYRo8+Hj4jeTcC2NmL6nqvv43ct8s/V9/G7FqMnH+oxMyuMG7+ZWWH62vg3tRTT91xdLDPXXF0ss4RcXSwz11xdLLPt8QM9+HDXzMza1ddX/GZmtkRu/GZmhXHjNzMrTPa/eSvpJOCXgEOprt//r8AVEXF907gGuZ5EdXG5x9dxdwPvjYi/axLTdlxHtUjJleU6LoNapObq+z4y122UeS1a20fGanLS/7wn4PXA7wHfPzDvUcCbgN9uEtcg168AV1NdTO7gejoGuBJ4eWpM23Ed1SIlV5bruAxqkZqr7/vIXLdR5rVobR+Z2Ftn2binnYDbxtz3qSZxDXKN/AYc1S+I3Zwa03ZcR7VIyZXlOi6DWqTm6vs+MtdtlHktWttHJk25H+OXpL1+l7eet6JhXGqukaKq7NhDYykxU8R1UYsl1yyDdUyNy7UWqTF930e62Ea51iIlbqpxDSbNdgJOAT4OvBj4EeCHgRfV805tEtcg1y8D11K9xVoFHAQ8jerXws5KjWk7rqNapOTKch2XQS1Sc/V9H5nrNsq8Fq3tI5Om7L/AJekQ4CXAOqq3PNuBD0TEvzaNa5DrSKoPTxbi/gF4X0R8tklM23Ed1SIlV5bruAxqkZqr7/vIXLdR5rVobR8ZJ/vGb2Zm7cr6dE5J+wBnAqdSjTWAB6ne+myK+n+tlLjUXHW+7E4v66IWDeKyW8e+16JBTK/3kY62UZa1SIlrMq6xUo8JdTEB76b6QfYVA/NWAGcAf9AkrkGuXE8v66IWKbmyXMdlUIvUXH3fR+a6jTKvRWv7yKSp8+Y+dnBw65j7bmoS1yBXrqeXdVGLlFxZruMyqEVqrr7vI3PdRpnXorV9ZNKU++mcIemg4Zn1vBUN41Jz5Xp6WRe1SInLdR1T43KtRWquvu8j895GbS9z3v0idVxjZX2MHzgHuE7SR4G7qI5nHQH8DPDqhnGpuV4PbJb0nqG4jcCbG8S0HddFLVLicl3HvtciNVff95F5b6Oca9HmPjJW9mf1SNoPOJE9T13aHBHfbRrXINfw6VL/AFwd40/12ium7biOapGSK8t1XAa1SM3V931krtso81q0to+Mk33jNzOzduV+jN/MzFrWi8Yv6Y8kXThw+0JJf7SUuAa5LpR05sDtMwcflxrTdlxHtUjJleU6LoNapObq+z4y122UeS1a20cWlXr6T5cT8ATgsQO3HwOsX0pcg1w/CfzYwO0NwE81jWk7rqNapOTKch2XQS1Sc/V9H5nrNsq8Fq3tI4tNPsZvZlaYXhzqAZB0oKQD2ohLzZWrLmox75q1Pa4+1yJV3/eRLrZRrrVIMc24sn7FL+lhwAXAccB9VKcuHQzcBLwuIr6dGtcg1/C1MAD+g/HX8dgrJjVXg2V2UYuUXFmu4zKoRWquvu8jc91GmdeitX1kotRjQl1MwAeBk0fMfw5wVZO4BrlyvXZIF7VIyZXlOi6DWqTm6vs+MtdtlHktWttHJk0zbdzTTsAtY+77RJO4BrlyvXZIF7VIyZXlOi6DWqTm6vs+MtdtlHktWttHJk25H+P/jqTDhmdKWrOEuNRcuV47pItapMTluo6pcbnWIjVX3/eReW+jtpc5736ROq6xcr9Wz0bgXZJ2sOd1KZ4AvKJhXGquXK8d0kUtUuJyXce+1yI1V9/3kXlvo5xr0eY+MlbWH+4ukPQUBq5LERF3LDUuMSbLa4d0UYuUuJzXsc+1SM3VRS26GH+uz4uO+kXSOi6mL43/AOBxVP+73RMR31xqXGquOvZAIKaNaTOui1o0iMtuHVPjcq1Fg5he7yMdbaMsa5ES12RcIx+fc+OX9CiqX6Q5gOp/PgGHA18HzouInalxDXItnC51LHB/HbfYqV6LxqTmarDMLmqRkivLdVwGtUjN1fd9ZK7bKPNatLaPTLTYp745TMD1wI+OmP9DwP9tEtcgV66nl3VRi5RcWa7jMqhFaq6+7yNz3UaZ16K1fWTSlBTU1cT4U5dSTwlLOQ2q76eXdVGLeZ9m2mhcy7wWqbn6vo+0vo16XItG+8ikKffTOe+XtGF4pqQfAf5fw7jUXLmeXtZFLVLicl3H1Lhca5Gaq+/7yLy3UdvLnHe/SB3XWLkf4z+U6njWwcDdVB9kPIFqBc+JiHtT4xrkWg9cDuxg79OlXhURX0iJSc3VYJld1CIlV5bruAxqkZqr7/vIXLdR5rVobR+ZJOvGv0DVJ9yPo/og456I+MZS4xrkyvX0si5qkZIry3VsMP5ca5Gaq+/7yFy30QyWOdd+kTquxfSi8ZuZWXtyP8ZvZmYt60Xjl3SzBn5WTNIVkm5eSlyDXNdLesfA7XdIur5pTNtxHdUiJVeW67gMapGaq+/7yFy3Uea1aG0fWYwP9SxC0sOBB6P+qrSklcDKiHigScws4nJUwjqmarMWOder7+NvUxf9YqrxuvGbmZUl96tzAiDpyVSfcgdwd0R8calxDXLleu2QLmqRkivLdWww/lxrkZqr7/vIXLfRDJY5136ROq7FZP2KX9KRwKXAl6nOWQVYT/WL8udExN2pcQ1yDV4L48t13LjreIyMSc3VYJld1CIlV5bruAxqkZqr7/vIXLdR5rVobR+ZKBK/4tvFRHVxorUj5j8G+HiTuAa5cr12SBe1SMmV5Toug1qk5ur7PjLXbZR5LVrbRyZNuZ/VszIi7hmeGdW3076vYVxqrgMj4vYRcZ8HDmwQ03ZcF7VIict1HVPjcq1Faq6+7yPz3kZtL3Pe/SJ1XGPl3vi/LOkFwzMl/QzwTw3jUnPleu2QLmqREpfrOqbG5VqL1Fx930fmvY3aXua8+0XquMbK/Rj/w4DfAZ4BLFxn+lDgFuD8qD/0SIlrkCvXa4d0UYuUXFmu4zKoRWquvu8jc91GmdeitX1kkqwb/yBVPzismHBNipS4xJgDyfDaIanjT41rK1fO65gal2MtUnO1Of4GMXMff67Pi476xcRxLSb70zkl7cvu36AMSXcD10fEfzaNS81VeyxwGNX/ut8FRhU3Jaa1uC5q0aBm2a1j32uRGtP3faSLbZRrLVLiGo5rpKyP8Ut6OvBx4MnAPwJfpfqEe7Oko5vENch1pKpfuf9V4PFUxf01SR+WtC41pu24jmqRkivLdVwGtUjN1fd9ZK7bKPNatLaPTBSJp/90MQG3AqtGzD+APX8FZ2Jcg1y5nl7WRS1ScmW5jsugFqm5+r6PzHUbZV6L1vaRSVPWr/hZ5PhVVB9g7NcwLjVXrqeXdVGLlLhc1zE1LtdapObq+z4y723U9jLn3S9SxzVW7o3/M5JeJWnFwgxJ+0l6JfCFhnGpuXI9vayLWqTE5bqOfa9Faq6+7yPz3kY516LNfWSsrM/qkbQPsBF4EbAv1afc/wl8CHh7RDyYGtcg16jTpVZTvcU6PxY/1WuPmNRcDZbZRS1ScmW5jsugFqm5+r6PzHUbZV6L1vaRSbJu/F1TpqeX5aqEdUzVZi1yrlffx9+mLvrFUrnxm5kVJvdj/GZm1jI3fjOzwmT/zd1UkvYHVgwfE5N0XETcPEXeJwBHAX+zcKqVpEMj4muJjz8Y+PeI+I6kpwCnAZ+NiCuH4lay+9t4u+rlfW5EviczxQ8wjBnnSRHx0YS4gyJi15j796pXPb9JzVpbx1k9L+oce9Ss7edF6nOijm1cs5RtPml71zFT7SN1fCvbfJ7bu56XtJ659YFl8Ypf0suArcD7Jb1+6O7frWMeLeltks6U9ChJN0r6tKQP1BtlIddlkg6r/34x8IdU1814h6T/WYfdKOmjkl5Uf8q+2LjOAG4ArpF0PPDbwAeBoyWdMxB3FPAxqm/rnQo8FPhfkv60/pAHpX/7b+J6Slo5ND0E+A1JK+on3kKuXx/4+4ckfRa4UtINkn6wQb2Sapa6jqlafl6k1Ky150XKc6JJzVLGn7K96/va3Eda2+Yp27uOa20fSVnPLvrARJH4Ta8uJuBNwBsXmwbibqb68gPA2cDLB+67sf53M/Bc4Bzg08Cx9fxnAlcOxA9+k24LsH/99wpg60LOeqO8gurr0+cCjxgx/tuAh1CdcnUtsL6evw+wZSBuM/Do+u8nApfVf58EXBXNvv03cT2BbwHbqJ6MN9bTN+rbNwzkGvz7w8C6+u8jgWtT65Vaswbr2MXzYmLN2nxepDwnGtYsZfwTt/cM9pGUb6u2tr1nsI+kPK/n3gcmTZ0397GDg5dQXYviGcBPDE8DcTcMPe5dwBMH72P3jr4P8KWh+MEn8icH/v7YUNzW4eXVG/MFwEeAS4A1o8YFnLewg4xY5vD4/3rwyVz/e+uYOn1q+Ak+bj2pfqrtauDHR+0Yi+ws148aY0q9UmvWYB27eF5MrFmbz4uU50TDmqWMf+L2nsE+MnH8bW7vGewjKc/rufeBSVPWh3oi4ipgR0R8IiK2Dk8DoV+X9MiB278OXKLqONiCB+qc/wXcO7SoRwz8fa2kt0p6KHCFpDMBJP0y8PkRY4yI+POIeB5wDdWrkwX3SXqxqgsrPRU4o871/ex5mO0BSU+q79sADH65pekPMExcz4j4EvAy4NmSXqfqan8xnHto3t9JevTA7YUrATaqV738xWqWtI5dPC8a1GzSOkLa8yLlOQHpNUsZf8r2hnb3kYnjb3l7Q7v7SMp6dtEHxkv9H6KrCdgnIeZI4KlD81ZR/a+7cNGlC4Gn138fMxD3IuDNQ499CdVbwL8Cvkb1IwevY/fbyGsSx35gvfEvrsfzy1S/q3kne76SWAP8Zb3MjwOH1/PXAs+q/34YcBHV29nr6ulm4DLggIFcyetZz3861dvPz4y4b0u9jE8DfwO8cCEv8JrUeqXWLHUdu3peTKpZm8+LEc+Jxw0/J5rWLGH8Sdu75X0k9XndyvaewT6S8rxeah/Ya5svZXuPmor6ApekfaPBNavrD3QeHhH3tziGhwPfi4j/mCLHQYz5Vt8S1vMA4Ccj4kNLHVOdp7V6TVrHNjWtV/2YVmo2kG/mz4uhWG/zDvaRgXxz3d57PTb3xq+hHx2g+kmyiT+sMCou11xN4pY7SU+ketX4+aH5PxsRfzaLuL7nGrpPVD/avWvEfVnuI9bMUp4Xw7I+xq98f3Bj7j8MUQJJ51K9ZX2lpD8auvuVs4hbBrmerOpUvoskHQHcTnU64w2qT7ms47LcR6yZ1OfFxDw5v+KXdCtw4vBbmfpt10ci4vjUuFxzNVjmVYz+cElUnyu9tH7MxLiMc306Io6t408FjoiIN9S3b4yIZ9Z/txa3DHLdSHVO95OA1wPPj4gvqfqQ8A0R8cI6Ltd9ZK7PsS6W2cU+Mknu39xd9EcHJCX9sMJAXK65UuPeCrwfOIHqibCYlLhcc/37wh8R8aeS3qjd37iMGcX1Pdd/RfWtzS9KuiCqs1GIiC9q4Ite5LuPzPs51sUyu9hHxovET4G7mIC3A6+i+gr2wrz9qN7SvKdJXK65GsZdu1ithuo2MS7HXMAfA4cN1eDPqL7AsmUWccsg14cH/r5x4G8Bt+S+j3TxfC1hH5k05X6oJ9cf3Jj7D0O0VtSMSXoM1bce/25g3kqq856fHhEvajtuGeR6JXBnRGyVtDZ2XyvnVKovEF1a385yH7FmUp8XE/Pk3PjNzKx9WZ/VY2Zm7etd45f0obbics3VxTJLyNXFMnPN1cUyc83VxTK7yDWod42f6uvPbcXlmquLZZaQq4tl5pqri2XmmquLZXaR67/1sfGPOx2qaVyuubpYZgm5ulhmrrm6WGauubpYZhe5dks9/SeXCXhmW3G55ur7+HPN1ffxuxauxTS59nhM0wfMe6K61scpDF2dD3hF07hcc/V9/Lnm6vv4266FJ08LU9aHeiS9BXgh8BTgz7XnT56d2iQu11x9H3+uufo+/rZrYTYo90s2HBcRPw6g6rcqL6L6aTPY87hWSlyuufo+/lxz9X38rdZC0k1UX6QatnAdmONSYvqeq+/jT801UddvOcZN7P1TZOew+8chRv427GJxuebq+/hzzdX38c+gFj8I3DVqP2sS0/dcfR9/aq5JU+6v+P9W0lERcSdARFwm6b2SHkX1SzRN4nLN1ffx55qr7+NvtRYR8Y+S/ooxUmL6nquLZXaRa6Jp/+eY5UR1fuqjR8w/Gbi4SVyuufo+/lxz9X38bdfCk6fBydfqMTMrTO6HesysJap+nOV04PHs/inE98aeV3qcGNP3XH0ff2qusbp+y+HJk6fZT8CvAFcDRwMH19MxwJXAy1Nj+p6r7+NPzTVpyvpQj6Q7gFVUv9e5x13seRrUxLhcc/V9/Lnm6vv4Z1CLbRGxgeEgScCn6mVOjOl7rr6PPzXXRF28+kidgAOAL7YRl2uuvo8/11x9H/8MarFtzH23pcb0PVffx5+aa9KU9Td3I+KbVG9hpo7LNVcXyywhVxfLzDVX7d2SrpV0jKRVkg6S9DRJ1wDvaxDT91x9H39qrrGyPtRjZu2RdCTwS8C6etZ24H0R8dkmMX3P1ffxp+Yax43fzKwwWR/qGUXSYW3F5Zqri2WWkKuLZeaaq4tl5pqri2V2kWsPqR8G5DIxdG2SaeJyzdX38eeaq+/jdy1ci2lyDU69e8VvZmbT6WPj1+SQ5Lhcc3WxzBJydbHMXHN1scxcc3WxzC5y7X5A/VahNyStiohvtBGXa64ulllCri6WmWuuLpaZa64ultlFrkG9e8WfuoIpcbnm6mKZJeTqYpm55KrP975Y0qWSHjsYJ2lTakzfc/V9/Km5Jmr6ocA8J2DluKlJXK65+j7+XHP1ffwzqMWHgBcDxwEfA35g4L4bUmP6nqvv40/NNWnK/eqctwHrgVvYfRwr6r8DeFaDuFxz9X38uebq+/jbrsWqiPgAgKTTgN+X9OKoO0aDmL7n6vv4U3ONl/o/RBcT8BDgzjbics3V9/Hnmqvv459BLYZ/ovGngN8dvC8lpu+5+j7+1FyTpqyP8UfEd4HL2ojLNVcXyywhVxfLzDVX7aOSThx43F8D/yTpfcCjG8T0PVffx5+aa6zendVjZu2RtBrYEBEfnSam77n6Pv7UXP8d78ZvZlaWrA/1SFor6WpJfybph4fu+4smcbnm6vv4c83V9/G7Fq7FNLkmSv0woIsJ2Ez1s2KPAT4MPHHgvhubxOWaq+/jzzVX38fvWrgW0+SaNHXe3McObs+iHAhcC6yob9/QJC7XXH0ff665+j5+18K1mCbXpCnrQz3Ag5IEEBH/BrwRePsS43LN1ffx55qr7+N3LWaTq+/jT801Xur/EF1MwC8APz007xTgOuCeJnG55ur7+HPN1ffxuxauxTS5Jk1JQblNwH7AkW3E5Zqr7+PPNVffx+9auBbT5FqYenM6p6QDgYjqx6Wniss1V9/Hn2uuvo/ftZhNrr6PPzXXyMfm3PglPQy4gOqCRPcBAg4GbgJeFxHfTo3LNVffx59rrr6P37VwLabJNVHqW4MuJuCDwMkj5j8HuKpJXK65+j7+XHP1ffyuhWsxTa5JU1JQVxNwy5j7PtEkLtdcfR9/rrn6Pn7XwrWYJtekKffTOb+jEb8gL2nNEuJyzdX38eeaq+/jdy1mk6vv40/NNVbux/jXA5cDO4C7qK4vfgTwBOBVEfGF1Lhcc/V9/Lnm6vv4XQvXYppck2Td+BdIegqwjuqDjO0RccdS43LN1ffx55qr7+N3LWaTq+/jT821qNRjQrlNwOFtxeWaq+/jzzVX38fvWrgW0+SKiLwbP/Bs4A7g08DLh+67oUlcrrn6Pv5cc/V9/K6FazFNrklTZ009aXDwSeAgqrczlzBwGhN7XqxoYlyuufo+/lxz9X38roVrMU2uSVPuZ/V8LyJ2RbVW5wEvkfQD9X3RMC7XXH0ff665+j5+18K1mCbXWLk3/gckPRyq7yUDrwI2qfqZsaZxuebq+/hzzdX38bsWrsU0ucaLxLcGXUzAscCPD817HPC/gS81ics1V9/Hn2uuvo/ftXAtpsk1aerF6ZxmZtae3A/1mJlZy9z4zcwK48ZvZlYYN36zIZIukPR5SZ+Q9BlJ53Y9JrM2ufGbjfbqiHgGcDxwjKRXLBYo6QRJa+c2MrMpufGbjRER3wF+CThLkhYJOwFYO68xmU3Ljd9sgoj4FnA38DRJH5V0k6T3SdpX0inALwK/L+l3JK2U9HZJfyXp45LWdTp4sxH263oAZj3xAPB9wAsi4ruS3gE8JyL+QtJTgS0RsUXSbwG3RcSvSjqc6trpJ3c4brO9uPGbpTkM+Bfg6ZKeAfwPqqskDjsZ+Lak0+rbB85pfGbJ3PjNJpD0WOBQ4EzgXuA9wNeorpA47D+An4yI/5zfCM2a8TF+szEkPQL4Q+ANwDHAO4CdwM+xu/H/F7Cy/vsvgTPqx+4raeNcB2yWwK/4zUZ7s6TfoNpH3hERV0n6N2Ar8C3gg8A5kq4BbgTeLen9wJuBSyRtpfoP4V3dDN9scb5Im5lZYXyox8ysMG78ZmaFceM3MyuMG7+ZWWHc+M3MCuPGb2ZWGDd+M7PCuPGbmRXm/wOvYl696It6cAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 잔차 시각화\n",
    "fitted_model.resid.plot(kind='bar')\n",
    "plt.hlines(y=0, xmin=0, xmax=42, colors='red') #잔차는 0을 기준으로 생성되기 때문에 y=0의 선을 그어준다. (기준선 역할)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2-1 오차항의 등분산성"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [],
   "source": [
    "# yhat은 model 형성을 하면서 feature를 넣고 도출된 회귀값들임\n",
    "yhat = fitted_model.fittedvalues\n",
    "subset = pd.concat([yhat, target, fitted_model.resid], axis=1)\n",
    "subset.columns = ['yhat', 'target', 'model_residual']\n",
    "\n",
    "# 따라서 yhat, 전체형법범죄 데이터의 차이는 잔차 값으로 남아있음. (컬럼 자체를 연산하지 않았지만, 연산 결과는 잔차와 동일함)\n",
    "# subset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 오차항의 등분산성을 확인하기 위한 플롯 형성\n",
    "fig = plt.figure(figsize= (10,8))\n",
    "\n",
    "# yhat값의 변화에 따라 오차항이 어떻게 변화하는지 육안으로 확인하는 부분 (등분산성은 yhat값과 무관하게 일정한 분산을 유지하는 형태로 도출됨)\n",
    "sns.residplot(x = \"yhat\", y = \"target\", \n",
    "              data = subset, scatter_kws = {'edgecolor':\"white\"},\n",
    "              lowess = True, line_kws = {\"color\":\"green\"})\n",
    "plt.title(\"Residuals Diagram\", fontsize=15, y=1.02)\n",
    "plt.xlabel(\"Fitted Values\")  \n",
    "plt.ylabel(\"Residuals\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Lag_range multiplier statistic :  0.271\n",
      "p-value :  0.873\n",
      "f-value :  0.125\n",
      "f p-value :  0.883\n"
     ]
    }
   ],
   "source": [
    "# 하나씩 시행해보기\n",
    "from statsmodels.compat import lzip\n",
    "import statsmodels.stats.api as sm\n",
    "\n",
    "#perform Bresuch-Pagan test\n",
    "names = ['Lag_range multiplier statistic', 'p-value','f-value', 'f p-value']\n",
    "test = sm.het_breuschpagan(fitted_model.resid, fitted_model.model.exog)\n",
    "\n",
    "# 독립변수, featue의 values (endog는 y값)\n",
    "# fitted_model.model.exog\n",
    "\n",
    "for i in lzip(names, test):\n",
    "    print(f'{i[0]} : {i[1] : .3f}')\n",
    "# which assumes independent and identically distributed\n",
    "# error terms, or the original Breusch-Pagan version which assumes\n",
    "# residuals are normally distributed.\n",
    "\n",
    "# The null hypothesis (H0): Homoscedasticity is present."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Lag_range multiplier statistic :  0.271\n",
      "p-value :  0.873\n",
      "f-value :  0.125\n",
      "f p-value :  0.883\n",
      "오차항이 등분산(Homoscedasticity)을 따른다는 귀무가설을 유의수준 5.0%에서 기각하지 못합니다.\n"
     ]
    }
   ],
   "source": [
    "def bp_test(fitted_model, sig_level):\n",
    "    # 라이브러리 호출\n",
    "    from statsmodels.compat import lzip\n",
    "    from statsmodels.stats.api import het_breuschpagan\n",
    "    import pandas as pd\n",
    "    import seaborn as sns\n",
    "    import matplotlib.pyplot as plt\n",
    "    \n",
    "    # 변수설정\n",
    "    yhat = fitted_model.fittedvalues\n",
    "    subset = pd.concat([yhat, target], axis=1)\n",
    "    subset.columns = ['yhat', 'target']\n",
    "    names = ['Lag_range multiplier statistic', 'p-value','f-value', 'f p-value']\n",
    "    test = het_breuschpagan(fitted_model.resid, fitted_model.model.exog)\n",
    "    \n",
    "    # 오차항의 등분산성을 확인하기 위한 플롯 형성\n",
    "    fig = plt.figure(figsize= (10,8))\n",
    "    sns.residplot(x = \"yhat\", y = \"target\", \n",
    "                data = subset, scatter_kws = {'edgecolor':\"white\"},\n",
    "                lowess = True, line_kws = {\"color\":\"green\"})\n",
    "    plt.title(\"Residuals Diagram\", fontsize=15, y=1.02)\n",
    "    plt.xlabel(\"Fitted Values\")  \n",
    "    plt.ylabel(\"Residuals\")\n",
    "    plt.show()\n",
    "\n",
    "    for i in lzip(names, test):\n",
    "        print(f'{i[0]} : {i[1] : .3f}')\n",
    "        \n",
    "    if test[1] > sig_level:\n",
    "        print(f'오차항이 등분산(Homoscedasticity)을 따른다는 귀무가설을 유의수준 {sig_level * 100}%에서 기각하지 못합니다.')\n",
    "    else:\n",
    "        print(f'오차항이 등분산(Homoscedasticity)을 따른다는 귀무가설을 유의수준 {sig_level * 100}%에서 기각합니다.')\n",
    "        print(f'오차항이 이분산(Heteroscedasticity)을 따르므로 선형회귀 모형의 가정에 어긋납니다.')\n",
    "\n",
    "bp_test(fitted_model, 0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'Test Statistic': 4.532638385708793, 'Test Statistic p-value': 0.33869493583624644, 'F-Statistic': 1.1157772150795675, 'F-Test p-value': 0.36747455079909225}\n"
     ]
    }
   ],
   "source": [
    "# 이분산 검정 추가 (white test)\n",
    "from statsmodels.stats.api import het_white\n",
    "\n",
    "labels = ['Test Statistic', 'Test Statistic p-value', 'F-Statistic', 'F-Test p-value']\n",
    "\n",
    "white_test = het_white(fitted_model.resid, fitted_model.model.exog)\n",
    "\n",
    "print(dict(zip(labels, white_test)))\n",
    "\n",
    "# Null (H0): Homoscedasticity is present (residuals are equally scattered)\n",
    "# Alternative (HA): Heteroscedasticity is present (residuals are not equally scattered)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2-2 오차항의 독립성"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.35995258335255625"
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 한줄씩 해보기\n",
    "# 더빈왓슨 통계 (DW-test)\n",
    "# 오차항의 독립성 검정으로 귀무가설은 '오차항이 독립적이다.'이다.\n",
    "# statsmodel에서 도출한 OLS summary에서도 DW 값이 나오는데 2에 인접하는 경우, 오차항의 상관관계가 없는 것으로 나온다.\n",
    "from statsmodels.stats.stattools import durbin_watson\n",
    "\n",
    "durbin_watson(resids=fitted_model.resid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.35995258335255625"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 한줄씩 해보기\n",
    "# 잔차 정의\n",
    "resids=fitted_model.resid\n",
    "# DW 통계치를 구하는 공식 적용 (시차 1 적용)\n",
    "diff_resids = np.diff(resids, 1, axis=0)\n",
    "dw = np.sum(diff_resids**2, axis=0) / np.sum(resids**2, axis=0)\n",
    "# dw 통계치 도출 완료\n",
    "dw"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### DW 검정에 대한 설명\n",
    "The test statistic is approximately equal to 2*(1-r) where ``r`` is the\n",
    "sample autocorrelation of the residuals. Thus, for r == 0, indicating no\n",
    "serial correlation, the test statistic equals 2. This statistic will\n",
    "always be between 0 and 4. The closer to 0 the statistic, the more\n",
    "evidence for positive serial correlation. The closer to 4, the more\n",
    "evidence for negative serial correlation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAFLCAYAAAD4RtogAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAABES0lEQVR4nO3dd5hdVb3G8e87MymkEYoBkV5FFClBioIB4QbuvSp2mhRBsIsUueBVECmiBLk0AQUpiuUqKEVAelGIhCIIgkQkFwLSQgiQPvPeP9bamT0n58ycOZme3+d55plz1l5773Xa/u1V9tqyTQghhNBdTf1dgBBCCINTBJAQQggNiQASQgihIRFAQgghNCQCSAghhIZEAAkhhNCQCCChR0k6UNLLkh6VNF3S/ZIOLi3/L0nHV6yzvaRXJTVXpF8t6X19VfbuqCxrHfknSzqjG/lvl7Rm90vW5XZHS7pU0mOS/ipp307yjpT0DUl/lPRQXuewijz3SfpLXl78fa4iz/45z98l3SbpnRXL98jbeVzSnyXtVFp2VF6vvP3f99T7EZZNS38XIAxJP7V9OICkzYArJL3N9onADcC5wLdL+T8AzALeA9yT12sBtgTurdy4pAOBdW2f0GgBJd0OHGj76U7yHA58FngKWAy0kX4zI4BbgO+X8p4I7ArMB1qBy21fJukuYHJe980q+zg+r7fA9q6Nvp5uOBO4x/YBklYEfi/pcdv3V8lbBLBdbc+TtBpwm6RHbP8pL3sLsIXt2dV2Junfgc8DO9ueJWlH4BpJ29l+QdK7gSnAZNvP5ODyO0kftP0YMAY4z/aZPfPyQ0+KGkjoVbYfBT4MfF3SWsBfgLXywauwMymgTC6lbQfcZ3txnxW2Ouc/AOW/FuC+inzH297B9i7ApUBzaZ0FpKDSWmX7J9resUrwWJzX7TGS1gC2sX0+gO3XSO/74dXy255u+2Tb8/LzF4AbgU26sdt9ge/YnpW3cRdwAfDlvPyTwDm2n8nL/wp8Aziue68u9IcIIKHX5bP8m4CPOk19cAuwC4CkFYCRwFWkQFLYjXSw6iDXPk4CvpCbYDaRNEzSabkZ5GFJJygZJun83Jz2uKSfSxon6SpgG+BGSb/upOhNwE+AD9v+iO2P2v4Q8G/AHRWvsTylw3voGGBMqoGUX8fXJN2ay3CXpDvz36k5S4f81Ui6WNKkrvKV7ADcWZF2G/D+elaWtC0wiVSLrNcIYGFF2oN5O/UsDwNYNGGFvvI32s9crycdhK8C3gf8yfbrkpA0PjeH7ArsV7kR25dIglITVj7oPm97m9z0dRmwFzAaGA4Ube57Ayva/kg9TVikWsBBwPtzn0czqYmqBXgR+EzlCpLGAOsCn5B0CbBxjW3/j+0f5HVuJNVgys11pusayLPAnC7ylK0F/KucYHtRDrQt1Wp7knYGfgCsBLyVVDv4V0W2/5Y0EVgZ+DNwjO1X8rLfAd+QNM32bEm7At8CJuTl1wDnSLre9kxJWwPfA8aXtr+TpN2ANYBXgW/b7hDAQ/+IABL6yjDaz6pvAr6TH38AuDU/vg34gKSbgJVs/7OrjSpFk8OA/8u1E0g1mhdIZ9vvBDa3/RfginoKmoNFC3Be/hPwKeDtwMmks+YRuZN7lu25pdW/A/zE9lXA8ZLupj0YLKlV2G7L+1oF2BrYiY79PV0GENvfquf1lKxAe3NcXWzfBmwBIGlVUo0M2vt/fkg6OTgmPz8G+BXpc8X25ZLGAzfnwH8n8Dng8rz8DkknAb+RNBx4ADiQVAsBuBt4HbjY9qs5UF0p6QO2n+zOawk9LwJI6CvvJNU8yAeCVyRtAOxIapICuBnYn3Tmf0ud210VeN32FtUW5iay7+X2/2tJZ6/zu9jmBsBXSH0XC4FFuUwLSAdIkWo2RW3nkbyvTwITbF8laQqpKevdeZvNLN2MNRr4JXAI6Sz9dtt/zoub6OE+EOA50sCEchmGkzrwu+xrsv1y7vT/ATmA2D6tItspkj4vacXcx4Lts4GzS/vcE3i0tN1fkt6HYvkWwBN52c2k70WRd5qky4A9gAgg/Sz6QEKvk7QR6Qz7ylLy9aQO1MW238hp9wLbk5qvlur/qOFlYKyk1Sv2ORbA9hW2J5M65dclBYZO2f677S/ZPjKXaVVSLWE7YCNgNjDF9tG2H1Ea7noc8DHg4LyNI23vCDyc+0cqhyhPAm4HLrT9W1KT248kFZ3LPR08IJ3N71KRthNL94t0ZlVS811Vpdrb3Fp5SB3ov+xk+Ve6WD6O9BmEfhYBJPSq3ORwLald/NnSohuAoykdvPJZ8NPAJ0jNWbUsAsqjuM4Czspn00jaCzhO0ickbZXzzM/rzSpvIzeB1Sr7V4B9SGfPewMfyWWeQ8dO9BWAmcDeVWo3Kv0v/95WBD5p+1f5tf+DFKCKs+0uayCSvl16fV2yPR34u6TP5PXHAycCZ+Tnq0r6Q+5HQtK6eR+j8/M1gNPoWJv4eh7ei6SRpCHav7C9qLTNj+THLbm/ap7tq0vbOFhSUx748EVSre38vGxPla4FkrQH6QRjyfqh/0QACb1hP0mPSHoC+C7wVdvnVOSZRmoWurUi/Sbgb7aXumai5FZSX8lfSR3z3yE1iTws6THSaK7TSMHoe5IeBh4mNeFcnLdxGfBbUnt9LWuQmqeesD3X9kLbz5GC3pgi+Nh+1falRb9GhXIQWNJkbPt3lX08tufZ/ltpva5qIevQMZDW4zPArpIeBf4InGn7obxsFPCOUjn/Req7uie/11cB37BdrrHMJl238VdgKjCDFGQL84C9JP0NKK7r+GRFmbYlfX5PkGqgk20XI7OeBo5QuojxYVINb/da152EvqW4oVQI1eVmsBOB95L6L4ofy2zStQ1/qrFqeRt7kkYiTQbeb/vYOvd9PnBSRa0thAElAkgIIYSGRBNWCCGEhvRrAJF0rComXqtYfrykaZIekPSxUnqzpPOUrjy+T2l+nRBCCH2oXwKIpLdJuhc4opM8ewCr2J5ImoLhC/lCJoBDgem2tyG1LZ8kaVhvlzuEEEK7fgkgtmfa3o6OozUqHUIawUMeGvkj0lBKSFcFn5uXzSJdU7B7rxU4hBDCUgbyleir5SGThSdJQ/wAWmwvqFi2YeUGJB1Kqq0wevTord/+9rd3qwAvvr6AF+YsfdHyauNGMmHsiG5tK4QQBqP777//ZdtvqbZsIAeQajobMrbUMtsXAhcCTJw40dOmTevWzm752wt8+ecPMndh+yzco4Y3c/beW/KBTVfr1rZCCGEwkjSj1rKBPArr5YrpKTYGpufHbcVVx1WW9ZhJm0xgi7XG05Qv5xo1vJkt1hrPpE0mdL5iCCEsBwZMAJE0QdK5paRLyX0kOVjsD/w8L7uKNAMrksbRvbmT6tbcJC4/eFs2nDCGNcevwNl7b8nlB29Lc1NvTFMUQgiDy0BqwhoLrF88yTOavkfSVFLz1CmlewycA/wwj+RqBY4q5t7pac1NYqVRw1lpFNFsFUIIJf0aQGxfUnr8D9IUzeXlVad9yMHikF4tXAghhE4NmCasEEIIg0sEkBBCCA2JABJCCKEhEUBCCCE0JAJICCGEhkQACSGE0JAIICGEEBoSASSEEEJDIoCEEEJoSASQEEIIDYkAEkIIoSERQEIIITQkAkgIIYSGRAAJIYTQkAggIYQQGhIBJIQQQkMigIQQQmhIBJAQQggN6Zdb2kpaDbgEWBmYAxxg+7mKPN8AdisljQNOtP1bSXeQ7pNeuM7293u31CGEEMr6657oU4ATbE+VNBE4HdinnMH2ycDJxXNJVwF356cjbG/XV4UNIYSwtD5vwpI0HljF9lQA29OAFXN6rXXeCbxg++U+KWQIIYQu9UcfyHrA9Iq0p3J6LYcDZ5aej5B0tqQ/SLpO0rurrSTpUEnTJE176aWXlqXMIYQQKvRHE5bo2H9RqJaGpNWB8bYfLyWfBEy1/aykjYD/lbSl7Q7bsH0hcCHAxIkTq24/hBBCY/ojgDwNbFiRtn5Or+ZLwLnlBNu/KT1+UtJMYCVgVo+VMoQQQqf6vAnL9ixgrqStACRtDrwCDJfUIVBIGgVsa/u2ivTtS483BUbn7YYQQugj/TUK63DgIkljgNeAg4CxpJpI2YHAZVXWnyTp20AzeRhwr5U0hBBCVf0SQGw/C0yuSH4B2KMi33k11j8VOLV3Stc7WtvM7U+8yKPPzWGzNcYxaZMJNDepv4sVQggN668ayHKltc18+qKpPPTMbOYtbGWF4c1ssdZ4Lj942wgiIYRBK6Yy6QO3P/EiDz0zm7kLWzEwd2ErDz0zm9ufeLG/ixZCCA2LANIHHn1uDvMWtnZIm7ewlceem9NPJQohhGUXAaQPbLbGOFYY3twhbYXhzbxjjXH9VKIQQlh2EUD6wKRNJrDFWuMpujtG5T6QSZtM6N+ChRDCMogA0geam8TlB2/LhhPGsOb4FTh77y2jAz2EMOjFKKw+0twkVho1nJVGwQc2Xa2/ixNCCMssaiAhhBAaEgEkhBBCQyKAhBBCaEgEkBBCCA2JABJCCKEhEUBCCCE0JAJICCGEhkQACSGE0JAIICGEEBoSASSEEEJDYiqTEEIYQGxjg/PjNoNJaWk5tNlLljunVa5XpEuw6pgRvVLWfgkgklYDLgFWJt/T3PZzFXlOB7YHFuWkmbb3zcuagbOBbfKyI2zf1QdFD6Hf1DpApEcsOcC058//ywef0rY6PqdDhsptVq5XbV1TWYA6XlNXy2uUp1qZKl//0vldNb2yHNVeY8d1yp9D+/tb9f0uL6f9cyv202E7OVj0tOEtGloBBJgCnGB7qqSJwOnAPhV5VgUOtv14lfUPBabb/oKklYGrJO1qe1GVvKEHtLWZxW2mzel/a/5b3NZGWxssbmvr8COC2j/Yen6sPan8w++sPHWVvUYhlzp4dpK3WFZ5YEkHj6UPKu5iWyH0l7oCiKS3AEcBawLKf7ZdedCvZ1vjgVVsTyVtZJqkFSWNtz27zs18Cpic158l6Xpgd+Ca7pZneTZ/UStz5i1i3qLWJQGh1WZxa3ugKAJHHMBCCJXqrYFcAfwIOK8H9rkeML0i7amc/mBF+uGS1gTGAOfZ/lVOb7G9oJTvSWDDyh1JOpRUW2HttdfugaIPbm1tZs78Rcyeu4jZ8xYtdZvdEELojnoDyPDSwXtZieqtFZVpPwbm2b5f0jjgakkP12jSqrY+ti8ELgSYOHHicnkOPW9hK7PnLWT23EW8Pn8xrb3RyBpCWC7VG0D+Jmlr2/f3wD6fZunawvo5fQnbd5cez5F0HfAO4HGgTdJw2wtzlo2BR3qgbANea5u5/YkXefS5OWy2xjgmbTKhw50NW9vMa/MWMXvuQl6bt4j5i9r6sbQhhKGs3gDSBNwm6U7gRdr7QD7T3R3mPou5kray/YCkzYFXgOGSzrX9RQBJWwOP2F6YayD/AeybN3MVcBhwdl62K6kjfkhrbTOfvmgqDz0zm3kLW1kh31v9gk9vzevzF+daxqJeGckRQgiV6gogtg8FDs0jntq60dldy+HARZLGAK8BBwFjSTWRwhrAKXnIrkijtmbmZecAP5R0L9AKHLU8jMC6/YkXeeiZ2czNfRdzF7bywIxXuexPM9hqnZX6uXQhhOVNvaOw1gZ+AryZn48kDbF9ppGd2n6WPIqq5AVgj1Kea6gxqioHi0Ma2fdg9teZry3V8b1gcRtPv/JmBJAQQp+rtwnrPOArth8FkLQZcAHw771VsNDONq+8uZDRI1oY3tLEgsXt/RrDW5pYd5XR/Vi6EMLyqt4AMroIHgC2H821kNCLisAx89V5zF3Yyqarj2PDCWN47Pk52DCipYkNJ4xhi7XG93dRQwjLoXoDyPOS9rB9PYCkycCM3ivW8q0ycBSamsRxe2zKMVc+zIJFrRy4w3pssdZ4mkqjsEIIoa/UG0A+D5wh6ZtAG2nI7Jd6rVTLqVqBo6ypSYwd2cLYkS3R7xFC6FedBhBJm+Wmq22BygsJ3wv8obcKtjypJ3CEEMJA01UNZHPgUdKsuJVMBJBlEoEjhDCYdRpAbP88P3yX7Y+Xl0nqqalNljsROEIIQ0FXTVjvByYBW0n6VmnRCGCjXizXkLWotY2Hn30tAkcIYdDrqgnraeB2YE/gjlJ6G8vB1CE9pahxvLlgMa12BI8QwpDQVRPWDGCGpA9VXnUuaSfgzt4s3GBX2VTVGjfVCCEMIfUO410k6TQ63lDqPcAGvVWwwSz6OEIIy4N6A8jPgPOBdYGTSLPi3tpLZRq0InCEEJYn9QaQNtv/K2kdYKHt/5J0O+kuhYF0e9gn/vV6BI4QwnKjqc58syStBvweOCaPzqo3+CwXFra2RfAIISxX6g0ghwGLbD8GXA18mHQPjxBCCMupmrUISadScZ9xacmkfQtIAeS4XitZCCGEAa2zZqgb+qwUIYQQBp2aAcT2kgsHJb0FOBFYy/Z/SjqQdJFhCCGE5VS9fSAXkYbyjsrPrwCObXSnklaTdL2kqZJukrRGlTzjJZ0n6TZJ90iaotyGJukXku6QdHv+m9JoWUIIITSm3pFUI23fXRzAbS+UNHwZ9jsFOMH2VEkTSdOi7FORZzRwge2/AEi6kHTP9N8DqwOTbc9fhjKEEEJYBvUGkBfy1CUGkLQHDd6RUNJ4YBXbUwFsT5O0oqTxtmcX+WzPBGaWVp0BLEvQCiGE0IPqbcL6IrAfsIGkPwIfovE7Eq4HTK9IeyqnVyVpR9I9Sa4tJZ+cm8FulrRzjfUOlTRN0rSXXnqpweKGEEKopt4ayGdsH9pD+xQVw4OzpdJyk9nXgLWBj9tenBd9F/in7SckrQ5cJ2k327M6bNC+ELgQYOLEiTGTYQgh9KB6ayCfkrRCD+3zaWDDirT1qT6q62Jghu3Dy/0dtm+w/UR+/C/gXtI8XaEb2trMAzNe5coHnuWBGa/S1hYxNoRQv3prIPcB90u6AnixSMxn+N1ie5akuZK2sv2ApM2BV4Dhks61/UUASR8Cptn+TeU2JG0L3Ge7TdJbga2Ao7pbluVZW5s55fq/Mf3FN1i4uI3hLU1sOGEMx+2xKU1N6noDIYTlXr0B5P7811MOBy6SNAZ4jXRV+1hSTaTwPmCSpE+U0n5t+xxgM+A7kpqBhcDnbM/rwfINeQ89M5vpL77BgsVtACxY3Mb0F9/goWdms9U6K/Vz6UIIg0G9AWQ/27v11E5tPwtMrkh+gTRMt8jz9U7Wv5jUvBUa9PQrb7IwB4/CwsVtPP3KmxFAQgh1qbcP5HVJu/RqSUKfWneV0Qxv6fjxD29pYt1VRvdTiUIIg029NZC3ADdLeorUByLAtnfotZKFXrXFWuPZcMIYHnt+DjaMyH0gW6w1vr+LFkIYJOoKILZ3BJC0cnrqV3u1VKHXNTWJ4/bYlGOufJgFi1o5cIf12GKt8dGBHkKoW10BRNLapD6Hufn5SOBg28/0YtlCL2tqEmNHtjB2ZEv0e4QQuq3eJqzzgK/afhRA0mbABcC/91bBQgghDGz1dqKPLoIHQH48sneKFEIIYTCoN4A8nydQBEDSZBqcTDGEEMLQUG8A+Tzw8XxfjruBPWl8MsUQQghDQL19IOsB59u+D0DSdsAGwMO9VbAQQggDW701kNOAv5eeP0G6KVQIIYTlVL0BpMn2a8WTfB3IsN4pUgghhMGg3gAyS9IWxRNJ7wHiDk0hhLAcq7cP5MvAjySNIE1jshA4pNdKFQaVtjbz0DOzefqVN1l3ldFxRXsIy4l6pzJ5UdLHgW2AZuCe0t0Bw3Is7isSwvKrriYsSe8G7gL2Aj4B3JFvBBWWc+X7ipiO9xUJIQxt9TZhTQE+Yvt5AElrAD8Ddu6tgoXBIe4rEsLyqzujsJ4vnth+DogbaIe4r0gIy7F6A8gTkj6vdocRFxEG2u8rotzdEfcVCWH5UW8A+SrpplJ3ArcBKwFHNrpTSatJul7SVEk35SaxavmOlzRN0gOSPlZKb5Z0nqT78t+OjZYlLJviviJvG78CbxkznK/sslF0oIewnKi3D+Trtk8ETiwSJB0LnNrgfqcAJ9ieKmkicDqwTzlDnrxxFdsT8/1HrpN0h+2XgUOB6ba/kG9ydZWkXW0varA8YRn05H1FYkhwCINHpwEk30hqPeCjku4sLRoB7EcDAUTSeFJgmApge5qkFSWNtz27lPUQ0vUn2J4v6UfA3sDZwKeAyXnZLEnXA7sD13S3PGHgiCHBIQwuXdVA1gMmkZqsyiOu2oDDGtznesD0irSncvqDpbTVcmd94Ulg+/y4xfaCimUbdrbTp156k09dcE9DBX7s+TkAna7f2mbeXNj5pTEzXpkLwInXPtppvq701HYGWplen7+YmbPn4Tw8Y8HiNh57fg7HXPkwY0fWW1kOIZQ1SYwZ0Tu/n063avsO0jUf99u+tof2KaqP4KpnVFdneZZaJulQUnMXY966QV2Fq+Ydbx3X8Lpl66wyakBtpye31RPbmb+odUnwKNiwYFFrQwGkCGo9Ubae2tZQLZNt/vnyXNpsVhs3kjEjmpEarzUOpNe2PJSpUfX+Kr8h6bj8+C3AusBNthu5pe3TLF1bWD+nl70saXXb/8rPN6a95tImabjthaVlj1TuyPaFwIUAEydO9C8P274yS4+ZM38Rj86c02vbXx48MONVzrr1SRaUrisZ0dLEgTus11DfSlEb+tZ/brbMZeupbQ3FMhVNjwtb27DhpdcXsOIKy9b0OFBe21Ao0/AWsfU6Kze8j199rvayukZh2d7e9g75byNSH8Q/GymM7VnAXElbAeQr2l8Bhks6t5T1UuDonGc4sD/w87zsKnITmqRxwK7AjY2UJwwcxZDgES1NiIEzJLitzbw+fzEvvb6AB2a8SltbXAJVVsxGUG56jNkIlg8NNYzZvlLS4cuw38OBiySNAV4DDgLGkmoixT6ukvQeSVNJzVOn2H4lLz4H+KGke4FW4KgYgTX4FUOCB9IorOLsuuibOevWJ6Njv8JAnY2gCPzzF7XywIxX+/27NBTVFUAkbVx+CmwLzG90p7afJY+iKnkB2KMi37E11l9EzAY8JDU1ia3WWalHhgP3xMGjs7PrmKolKWYjKDc99vdsBBH4+0a9FxIeW/o7htQHslcvlSmEZVI+eLz8xkLOuvVJTrn+bw01PXV2dj0U9ETzXE83PfZEmaJZrW/UO537QZVpkr4G/KDHSxTCMurJWsNAPLuGnqlh9dRZek82PfZUmQZqs9pQU+907u+TdLekJyU9JOkZYLdeLlsIDenJWsNA7NjvqRpWT56lF02PH91qTbZaZ6WGm4l6qkwxyWffqLcJ6wzSyKsbge2AzwN/7a1ChbAsevLgUZxdf2WXjfj41msOiLm+euogOxCb53qqTAOxWW0oqncU1mzbMyU9DGxu+1pJDU+mGEJvKg4elVOiNHrwGGgd+z3VPDMQm+d6qkwDsVmtpw2EUWb1BpC/SNoM+C3wU0l/BOKquTAgDfXhwD11kO3pQNsTerJMPRX4B+JIvIES1OrtRD+6eCzps8AWwGm9VKYQlllPHTx6Sk8ehHrqIDsQA+1ALNNA7JAfKEGt2xcS2p4BzOiFsoQwZPXkQagnD7IDLdDCwCvTQGzqGyhBLaY4DaEP9PRBaKAdZIeynmxW66l+i4ES1CKAhNAHBmJ/Q6hPT9X4erLfYqB8n+qdymQtANvPSPpP4H3A6fnugCGELgzEtv1Qv56o8fVkv8VA+T7VWwO5FNg337v8i8DlwE+AD/ZWwUIYaqLZafnW0/0WA+H7VO+FhCNsPw/8O3CO7SuAuKQzhBDqNBSvjq83gCyStAtwMHCzpBbqu4NgCCEEBua0OMuq3iasg4EvAcfaXiBpB+DMXitVCCEMMQOl36In1QwgkppstwHY/gfwtWKZ7T/1QdlCCGFIGQj9Fj2psyaseyV9VtKwPitNCCGEQaOzALI9MB04oo/KEkIIYRCp2YRluxW4Lf+FEEIIHdR7Q6n1Kp5L0n81skNJq0m6XtJUSTfla0uq5Rsv6TxJt0m6R9IUScrLfiHpDkm3578pjZQlhBBC4+odxnuZpA0BJK0D3AI02jcyBTjB9rake6yfXiPfaOAC2zvb3h4YC+yRl60OTLY9Kf/FvUlCCKGP1TuMdy9SEPkD8CHgS7Yf7O7OJI0HVrE9FcD2NEkrShpve3Y5r+2ZwMxS0gxgeHf3GUIIoXfUrIFIGl78AS8BnwY+BhwJPJrTu2s9Usd82VM5vSZJO5I69a8tJZ+cm8JulrRzjfUOlTRN0rSXXnqpgeKGEEKopbMayI2kq82Lq1wMvAmcktMM7NLN/RXrVap6VXvu8/gasDbwcduL86LvAv+0/YSk1YHrJO1me1aHjdoXAhcCTJw4Ma6cDyGEHtTZKKyqZ/XdIekIUpMXwB3A/wAbVmRbH3i6xiYuBq61fUZF2W4oPf6XpHuBdYFZhBBC6BP1dqI3xPYZpY7u43MNYa6krQAkbQ68Ynu2pAmSzi3WlfQhYJrt31RuV9K2kpry47cCWwF/683XEkIIoaP+uKHU4cBFksYArwEH5fSxpNpI4X3AJEmfKKX92vY5wGbAdyQ1AwuBz9me1+slDyGEsES9N5SaaHtaRdrWtu/v7g5tPwtMrpL+D9qH6WL7651s42JS81YIIYR+Um8T1veqpJ3dkwUJIYQwuHRaA5F0AHAgsIWkW4tkoBn4c+8WLYQQwkDWaQCxfSlwqaRzbX+xj8o0KA1vbmJ4S9NSt6wMIYShqq4mrCJ4SBqWLy4cIenMXi3ZIDNyWDNbrjWe9VZd+raVIYQwFNU7meLHJD0CPAPcRbreYvDeyLeXNDWJ1VccGYEkhLBcqPcIdyywHfAL4AOkIbbNvVWowS4CSQhheVDvkW2W7TeBB4Bt8kSKG/ResYaGCCQhhKGs3gsJb5C0GfBb4HeS9idNghjqUASSCWNH8OLrC5g5e150tocQBr26Akh5LipJHyTVPh7urUINVRFIQghDSbenMrH9BvCXXijLciMCSQhhKOiPubBCFoEkhDCYRQAZACKQhBAGo3onU/wxS4/YmksalXVZ6UZPYRlEIAkhDCb1jisdBVwHnAacC8wj3a98PHB+r5RsORbDf0MIg0G9TVirV9zY6X5JN9vetTTJYuhhUSMJIQxk9Z7ariBpQvFE0nhSrQRq3M889JyokYQQBqJ6ayBfBq6U9CjQCmwJHC3pA6TpTUIfKNdIXn5jAa/OXcRr8xbR2hYxPITQ9+q9kHCapB2BTYFhwGPAGNuvArf0YvlCFU1NYsK4kUwYNxLbzJm/mNfmLmL2vIW8uaC1v4sXQlhO1DsKS6Tb0K5FuqHUdsDXgLd3d4eSVgMuAVYG5gAH2H6uSr7Tge2BRTlppu1987Jm0h0Rt8nLjrB9V3fLMhRIYsUVhrHiCsNYm1EsWNzKa/MW5YCyiMWtUTsJIfSOepuwLgHeADYHfg58EjirwX1OAU6wPVXSROB0YJ8q+VYFDrb9eJVlhwLTbX9B0srAVZJ2tb2oSt7lyoiWZiaMbWbC2FQ7eWPBYmbnpq43FizGEU9CCD2k3t7YdfNNpW4hTag4Gfhod3eWO99XsT0VUtMYsGJO745PkYYTY3sWcD2we3fLM9RJYuzIYay18ije+bYV2XqdldhotTG8ZewIhreov4sXQhjk6q2BvClpGHAnsBdwHjCugf2tB0yvSHsqpz9YJf/hktYExgDn2f5VTm+xvaCU70lgw8qVJR1Kqq2w9tprN1DcoWVYcxOrjhnBqmNGAPDmgsXMnreI+YtaaW3zkr/FpcfRQR9CqKXeAHIcsI7tWyX9J6kmckoD+xPVh/1WS/sxMM/2/ZLGAVdLerhGk1bVbdi+ELgQYOLEiXEkrDB6RAujR3T9FVjc2karOwaVJX82i1tNW15efpOXbi5z1WWdr7PsXGu/7nx5p2lVvrLV81XbXlo75Tc2S553XBbCwFbz6CHp3yqSJkhaH7gBuJE6rv+QdATwofz0DuB/WLqmsD7wdOW6tu8uPZ4j6TrgHcDjQJuk4bYX5iwbA490VZ7QmJbmppg0rY/Z5cDSHlScAw6l58VjaM9fPC4vo0Z+OllnSaCsyF8tT70BtNbrrZ5eO61a4K98PZVl6Li92icOtdYp3v9qn0WH5SwfJwqdHRe272JdA3/oNEO6j8gZ5TRJcyVtZfsBSZsDr9ienS9UPD73tSBpa+AR2wtzDeQ/gH3zZq4CDgPOzst2JXXGhzAkSEJLuqmiv2ooK58stC153DE4tS0JOO4YhCqCVVtOKJ9wqBe/PzUDiO1v99I+DwcukjQGeA04KKePJdVGCmsAp+QhuyKN3JqZl50D/FDSvaQLG4+KEVghhMGofLLQPMhOFlSr6jjUTJw40dOmTevvYoQQwqAi6X7bE6sti0mVQgghNCQCSAghhIZEAAkhhNCQCCAhhBAaEgEkhBBCQyKAhBBCaEgEkBBCCA2JABJCCKEhEUBCCCE0JAJICCGEhkQACSGE0JAIICGEEBoSASSEEEJDIoCEEEJoSASQEEIIDYkAEkIIoSERQEIIITQkAkgIIYSG1Lwnem+RtBpwCbAyMAc4wPZzVfJ9A9itlDQOONH2byXdQbpvfOE629/vvVKHEEKo1OcBBJgCnGB7qqSJwOnAPpWZbJ8MnFw8l3QVcHd+OsL2dn1R2BBCCNX1aROWpPHAKranAtieBqyY0ztb753AC7Zf7vVChhBCqEtf94GsB0yvSHsqp3fmcODM0vMRks6W9AdJ10l6d7WVJB0qaZqkaS+99FKjZQ4hhFBFXzdhiY59F4VqaWkFaXVgvO3HS8knAVNtPytpI+B/JW1pu8N2bF8IXAgwceLEmvsIIYTQfb0aQCQdAXwoP70D+B9gw4ps6wNPd7KZLwHnlhNs/6b0+ElJM4GVgFnLWOQQQgh16tUmLNtn2J6U/463PQuYK2krAEmbA6/Yni1pgqQOgULSKGBb27dVpG9ferwpMDpvO4QQQh/pj1FYhwMXSRoDvAYclNPHkmojZQcCl1XZxiRJ3waayUOBe6WkIYQQalJFt8GQNXHiRE+bNq2/ixFCCIOKpPttT6y2LK5EDyGE0JAIICGEEBoSASSEEEJDIoCEEEJoSASQEEIIDYkAEkIIoSERQEIIITQkAkgIIYSGRAAJIYTQkAggIYQQGhIBJIQQQkMigIQQQmhIBJAQQggNiQASQgihIRFAQgghNCQCSAghhIZEAAkhhNCQCCAhhBAa0m8BRNKxkj7XRZ7jJU2T9ICkj5XSmyWdJ+m+/Ldj75c4hBBCWZ8HEElvk3QvcEQX+fYAVsn34t0B+IKkVfPiQ4HptrcBJgMnSRrWm+UOIYTQUZ8HENszbW8HHN1F1kOA7+Z15gM/AvbOyz4FnJuXzQKuB3bvlQKHEEKoqqW/C9CJ1Ww/V3r+JLB9ftxie0HFsg0rNyDpUFJtBeANSU8sQ3lWBV5ehvX7S5S7b0W5+1aUu/etU2vBQA4g1bg7y2xfCFzYEzuWNC03pw0qUe6+FeXuW1Hu/tWrTViSjpB0e/77djdXf1nS6qXnGwPT8+M2ScNrLAshhNAHejWA2D7D9qT8d3xneSVNkHRuKelScj9JDhb7Az/Py64CDsvLxgG7Ajf2dPlDCCHUNpCasMYC6xdPbF8l6T2SppKap06x/UpefA7wwzyaqxU4yvaiXi5fjzSF9YMod9+KcvetKHc/kt1Zt0IIIYRQXVyJHkIIoSERQEIIITQkAkgIIYSGDKRO9H4naTXgEmBlYA5wQMXFjEW+44EPkgLwybZ/05flrCjLDsCxwBhgBPBt2zdW5DmddBFmMdBgpu19+7SgFfJFnc+Xki6yfXmNvAPp/W4GbqlIXs32phX5fgG8lfbrk+63fWQfFLEqSccCr9o+Pz/fhNSRuwLwDOm7/kaV9ZqBs4FtctIRtu/qm1JXLfd/Al8ifdebcnnur7Jev77/5XJLWpM0SvSlUpaTbd9UZb1+fb+7zXb85T/gp8C2+fFE4IoqefYAzsqPR5IOJqv2Y5l3BMbmxysBD1fJcwnw9v5+fyvKdG+d+QbU+12lfGvV+J7cDowcAOV7G3Av6eD1uVL6DcD6+fGHge/XWP/zpIMYpBOrO4Bh/VjuDwAj8uN1gbtrrN8v73+1cudy/qLO9fvl/W70L5qwMknjSZM3TgWwPQ1YMaeXdTZHV5+zfZft1/PT2cA8Seqv8vSCAfV+V/EV0hnjgOQqc89Jehfwf7afynl+R5qwtJp+mXeuWrlz+i1un8boX3Q+O0Wfq1XubhhU8/xFAGm3Hktfzf5UTi+rNkfXUvNw9TVJTcDppKagaj+qwyVdm2cF+GQfF6+aVSVdJOkWSb+WtG6NfAPy/QaQNBZ4p+17amQ5WdL1km6WtHNflq0LGwKV88LNkrRSlbx1zTvX1ySNAH4MnNpJtoH0/m8g6aeSbpN0SWlm8UoD8v2uJfpA2onqZzP1nOH061mQpAnA/5Cqyb+rkuXHwDzb9+cr96+W9LDtx/u0oB19G7jO9ixJ2wE/Aer9kQ+Us85DgItqLPsu8E/bT+Qpea6TtFs+q+xvy/Jd706+XiFpA+D7wBTbf6yRbSC9/y8DU4BrbL8p6aOk32u9/ZAD5fu+lKiBtHuapSP9+jm9rLM5uvqcpPVJB9+jawQPbN/t3NFoew5wHfCOvitl1TJdXvyYbd9L6sytZkC934Xc2flB0rQ6S7F9g+0n8uN/kdrF1+2zAnbuKdL7WLaK7dlV8g6oeeckbQucBhzcSfAYUO+/7Tds/8L2m/n5lcAGNbIPqPe7KxFAsnwwmytpKwBJmwOvAMO7MUdXfzgB+IztZ4uEynnFJG1dfClzDeQ/gKl9XdCyPHqsePx+0kig7s6J1p8+DvzOdiukfgVJ3ywWSto2Nysi6a3AVsDf+qWkFWw/BGwsaR0ASbsDd+XHHV4HA2/euWOBT9t+tZw4kN9/SeMkbVZ6vg9wT3480N/vTkUTVkeHAxdJGgO8BhxE9+bo6g87AL+s6Dc/jFKZgTWAU/JZs4ATbM/suyJW9ak8PHcY8BxQ3N54oL/fhYOBj5aer0wakVXYDPhOfs8XkkbkzOvD8nXlq8BP84CLmcBncnrl6+iPeeeqkjSK9H2/vuL7/p8M7Pe/FThS0tuA4cBjwFF52YB9v+sRc2GFEEJoSDRhhRBCaEgEkBBCCA2JABJCCKEhEUBCCCE0JAJICCGEhkQACaFBkk7I11D0Zxnu7c/9h+VbBJAQBjhJ4yXt2d/lCKFSBJAQBr7xwJ79XIYQlhIBJIQeJmnfPAPsNElfLaUfKOl+SXdK+p2kAyvWa5H0k7zuPZI2y9O3/ALYPc+kPJwqJO2SZ56tNadYCD0uAkgIPe9h27sC2wH7S1oxT1e/D7C97Z3Ic39VaAK+m9f9KnC87YXAXsANtifl5x1I2gb4OvDxATZdShjiYi6sEHre9HzPlW1Jd4lcnTQp3qWlAPDXypVsL5S0QNJXgHfRcY6kWtYArgX2LGZ7DaGvRA0khB6UJye8GlgMfAeYRprAclhOK4ypsu62wJmkmXG/nterxweB0yUNa7jgITQgAkgIPWtFYHi+58NYYBdSILgZ2E9SU55Vdq8q625PmiL+wYrlbaRZXKt5zvafgV8CJ/bQawihLhFAQlg238ud27dLuh0YB/xR0p+As0j3cP+u7b8CfwD+CPwm/59bsa0rgE9LugMYBYyStBdpuvu1O+tEJ92XfaKk9/Xw6wuhppjOPYQ+IGllYF3bD+RmriuBY2z/vZ+LFkLDogYSQt94kzQi6xZSH8ftETzCYBc1kBBCCA2JGkgIIYSGRAAJIYTQkAggIYQQGhIBJIQQQkMigIQQQmhIBJAQQggNiQASQgihIRFAQgghNCQCSAghhIZEAAkhhNCQCCAhhBAaEgEkhBBCQyKAhBBCaEgEkBBCCA2JABJCCKEhEUBCCCE0JAJICCGEhkQACSGE0JAIICGEEBoSASSEEEJDIoCEEEJoSASQEEIIDYkAEkIIoSERQEIIITQkAkgIIYSGRAAJIYTQkAggdZJ0gqSHJd0p6T5JR+T0FSTdW5H355L2Kj1fUdI13dhXs6T4bGqI96Z+yvq7HGFoih9i93zd9k7AjsB2kr5gex7wpqSVS/lWASaVnu8I3Fltg5LOl3SPpJsk3SjpZuBmYNtSnhZJLRXr/Tz//6mk0RXL3i7pnZLGlNa/uqsXJ+kYSTvVWDZM0ncl3SbpLkmXSRrRybaOlLSLpP+RtGGNPKdIulnStZJ+J+kaSddJuqSUZwtJ10u6Q9LTkq4HPp6X/UzSyIr3aVL+WzenrSrpp1299pz3Ckmr15m3JR+bb+4iT1N+fJSkf6tYvqGkMyvSru9iv8r/r6qx/LqKpOOAD3S2zbzeyZLeV2PZapLOkfT7fCI1Mqe/W9KJNdb5iKQzutpvxTrrSrqojnxn1fpOhb7V0nWWUMn2fEkHA3+W9EPgNmAn4Lf5iz0N2KS0yiTgF51s8jfAc0Ab0AyMydsoHA18QFIrMNp25Q99bsXzLYGRwCvAG7YXl89CJd0CvAEYGAecYPtOYAG1TyqGA9fb/q+8jeOAg4HzJI0HzgdWBi60/WtAQGveXluNbX7L9uJyQi7nlcVz2w8Be0jaGTjK9n+UshtYnNcryj0qp5e/20v2n/Pdlf9M+mzea7sNWFRsr6JMbwF+RXqftwTuz6/rM3mdIt+7gdPy6xbwVuAM4PIa234nsH5F2pul7a0M/BZ4Pb+GscA9wLE1yrkWsLWkltL7amB+Kc8WwA9I37P7gSNsm/TZ16qpnJP/pgKHAcdI+l9gw/LrL+1jfWBv4BlJu9u+oWL5T0if/TMVq87L5S3yXQ6sBSzMSa/a/lTOs6BGWUMfigDSINuvS5oObAzcAnyK9GPfiXRwGiNpTdvPAu8GjqmxqSbgQWA66SDRRPpxLzlA2D4VODUf/H5UJOf/i/IBAEkHAnvlZcOAfSVNtf0N0kGt2N6SM1JJ3wVm1fF63wTuKCXNAN6SHx8FXADcDvwyn0U35X2aGgGkMnhkawL/rJJ+EDA+HyTfTwpe7ywtPxbYmXSQN9AiaThwJKWDku02SS+VAuFvgYMkfRx4F1UCqO2X8raRdJ3tDxbLyoHZ9l+A3UvLjgP+np8uovQZZJ8Gnpe0me1Hc9qS/dueBewk6V3AP4DVgX2KxVXeo+8CfwMOkDQD+AawHun7WDgD+ITtlyVNAT4IXE3nAWQYMC2fON0DHAisRjphKB/wdwY+mZ9+zvYsSYdL+jzwZ+DUHKhNe1AoW1RRhvG2J1XJp07KGvpQNGEtmzeAVYH7gK1z2nuBP5EOtjtLGge8brvy4FF4gVTDOIv0Qz6LdLb3ySp5JwFzcrPJFjmtvN3LbO8OfBl4xfauOXhA9QMOudyP1lhWlaTNSAf0i3PSZrZvy4FsGvAO2n/gZukDZ63tNgPbA49IGltqrjkAeA04BLgUuNH2zsCNxbq2TwY+BtwEPAR8z/ZupJpd5WsvBzQD/wscCvyRTg5MklYCdsj/C8Nq5G0hBZ0/l/ZTrgkdA9wAfB04LQdGgNVzs957Sps7AFgJWIEqB97clHYqMBOYDHwUaMrv0Y9JJyRIGks6i385r3o56fsK7Scv1ZwGXCPpTuAE4GTbtwH30vEkdAapZvH5HPywfSbpc/l1Dh7QfnJRyXR8/2vVMlS8ptC/ogaybNYBXs5NRLNyk8NKtl+TdAfphzeLKv0fkoaRmkOuBK4ifRZTgO+QDhTDJG1p+8Gcvxn4PLCv7YWSfpY3VT5jLX6gewGbSBplu2jeWurAKGlf4HdFDSbnmSLpn7Y/Xu0F53V2Az5ue05ObpPUlPc/HDgPWJF0QDZVgpekc0j9PHPzXzPpIFacXe4DfEzS/qRmjMNttyoNXjgB+CKpllaurf0YOBV4HDghv8cPAbtJuhX4mO1XgfflZrwmYFNgM9KBdJ1q71PJ0aQmnGNJB36A1STdCBxt++FS3q8APy69t0veB0lfAkbZPi0/Pzjn/wbwou09K/bbSgpUI4F/SloFWD83nxbNYzNtn5O3tycpiBSK74jpeOBdTPtnU3nwXsL2PcAuVRY9B/wu7/PtwNeAuZIW5jIXQWkYMFrSDbavzmkteT2R3vetgacrylCr6VPAZZJ+kwNU6CcRQBokaU1gAu1NFLcB+5LOAslNBBNItYZqnbjNwAaks6xFpDPLE0k1muJHN7+U/wekmslO+ex1rdJ2yuXaF1iXVAv5qaT9bb9BxdmlUmf5HqRmlCXJwJG2b6/xmk8FnrN9YMWiaaRmoGuAXUkB5rBq5SvY/pKk/YDXbHc2Qu0cSWsA2wD35j6RL+ZlV9PxgLMa8KTtuZKeItWEHgJusH1wKd9bq9QI78kHwaoHUUmHALNt/0rSZpI+Z/t80gF/ckXeL5ICxC9Lyc2lbZ+Xm9ImAp8F1gacg+r9pe18B9iTdCCdROojeCO/zudsf760/fskbUJ63zcGmiVNAn5N+m5i+w1JK0ta2/b/kYJMUUNqqvXac1mOtv19SUcDP7I9G/hZrvFi+3HgsBwMLy03T0o6zPYF5c0BV0laTPruP0bqX3mtogwj8knAKFL/z1q0n2DsZ3tGrfKGvhEBpAG5KeDHpKp8cQZ3C3A9cEQp6wzSwfTrVLA9H/i5pK1JgWMk6UdsYDZpxNd0pZFUl5CaAIo+iJtLNZAlB2hJh5NGgB2Wz9ZH5PKcSHszxgjSAXgSsHep/F295ncDLbbPrrJ4CnASKSAdb3t2rjGJLg5MVOmErWJ1UjD8B+k725y3uwmpBlEE2tOAGyS1kd7D/Ug1osqmmQmS9rP9/Yr0YdXKKmlX4G22vw1g+3hJ/yVpc0pNMZJ2IPV1/cb2uRWbaSrKkYPHGOBMUrPcUznP5sAFks6z/Yrtb0r6FjC5Skd0ZdAaQRqo8UXgAVLQ2RA4F3gZ+EvO+mXgCkmLgIdJ/XYdylfDAZL+g/Sen680cGJrSSvlWl3hI8CWkubl58NI/ULlAHJQtf4vSavS8f0fCTxCCiyvkGqW19I+QKNY7wrgANv1fJdCD4oA0j3fy2f/LaSzyJ+Xlv01p99dSrsdWL2Lg/QxwFdtTy8SJO0OfI406ugNSYfkM76y4se+JIBUVudtF0OCy/nGkH6Me5aavJbsupNyvheYLGmbUtrdtv/b9kKWDpJFACmfeVdzuqSjaG/yUC777qU8bcAltk/qUNg0PHfJd9j2dcB1FXkmsPSBcRgdR8kVvgrMqUyseB+LtO/m7ZdrWI+Tzoxfr7LtYRXlGEdq3vxHceCTdD+ps3wU6TPCtiV9mtRfUlY5fHY4qSnwgXxyAvCYpIdJnd1Fuf8CVBuu21Wgf9D2p5WG2Zr0Xu0HfJOOJ00iNTeWayA3ljdULXhUK0Nlza60vU9UlPUdpO9BBJA+FgGkTrZPILW911pu0plyOe3XpCaEzpxC6ncYTXs/wOukkU3FdmZXWa9o9rGkYXWcfTXnbb1C6oiupuYBxPZ5pL6Net1JqoHVDCC2f0r15r1KzcD+kt5Le3/HCNKoqa46UysP3JDe43/P/SDljt3hwPdJTWP1Kgewzkazdehst/1cPnO+oRgskLd1jZce3rqLpNtyWZ3LKkmTc/AuRgX+ALhOpYFhpOarO+jaC6Qz/Vq2UbrGZHPSiMNFtm9UuhZkW9tTc75m4CZJc0knBSvkderRnZFV5UCzRTfWCz1IdbZghNBvlIYvjwDm19vkFnqWJBXvvdIIs9Zqn0X+rJo6qWWEISQCSAghhIbEdSCDUEW7e+gmxVxaA0ZPfBZ5pFa19B692LCntzcUxA+pG5TmBLpe0lSluavW6CL/dUpzMV1WY/nGkm7J27w6/10r6Q95CGaRb0re1rNK1zNcldM3UrqSvLzNjZTngyqlnSVp4zpe39sldaefA0kH5E7eevNLac6vz3WRb0jNpSVppKQfKE3GeZukW/Pjfarlr7L+wSpN0FmxTJXPlSbkPF/SBjXWac5NUUiaKOmUKnmurXh+Zmffo6Ickn6kNKKqcvmZkjYqPX8vSw8GqLbd3ST9dydZbpBU/L2Uvw+/B/67tI3KYexN+bv7SUnfVxplWCz7RP6O/rqU9ilg/67KmvPeWkeejyoNeR7UohO9e6aQ5o2aqjSG/3Ty1BL5QPAZ4HngC3kkTtHpWcuTwL9VXpOgNNJryTUgto/M6XcAx9i+r1hEacqT/AMeRuq4LGuqyPc10siVl0jTkfzN9hmka1KqXrwl6Z2kYafNeb/H276LNPKnO9+jr1Ka7ylvezxDeC6t7CPA806TcRbbWoHU6f2/pZFYzaTrfTbNr/+6/NlUzndW9jOli1g3AP6PdL3ImaT3cMmV65Juys9FGjAwgzQlzFLlVrpY8T3qOK/W4ort/Yz0mS0ivV8jSRccNlFxxXx+3yaThqQXKufpWpk0PH5l0jDsz+SBCZ1Ns7Jkap78G1wZ+KbzBbglJ+ff7DtJ0708Srr2pPgOtOZtfIV0vcnU9FTfJ33WxbQ8RVkPBBbngSCVyvOZfY302S/M+xlBmies1nQug0oEkDrlg9wqxWgT29OUpmkfT/oy7AP8G7AdcDjpivLi6uNac0HVmuZjI9IBtbz/t5O+fAeQLhq7CRhPns5D0ttIVyUvyH9N+cdQXAVfDmTjSMOQH5S0FfARpaGWw0k/rmq+Aexj+8V84PsF7QfRnSW9YvsPNdYtXsO7SJMHXk062BSG9Fxa2QPAp5Su35hB+iy3AJ6pGEH3b8BDxUWCki5WqukuIH0+1d6H4iTmfOAkp/nXkPRJOo5W2q14rHTNyo75abV5ur5I+j4dTMdrOMozH+yby9Zs+xlJF5fyVZ44HQz8CzhC6XqlK0nf39+U8hwPnG37NkkfJNUgjqCLACJpU9LJ279yufeTdBBpSPetthfZPjbnnWp7l/y4mDeu/PmeJekLpM+mGfhpLs8ngKMlvcf2l0jfyVojH8vv5brAgbafKmfIJzqDvkksmrDqtx5pwsOyp3L6JsCfbLfZ/hPtwxZF9+aCklJ77spAq9qnzV4R+CHpSvcRkg7KB4O9i3Vtz8w/jFNJF4jdRLpQ8NQiS2lX5fIUZ0KHkA6itb7UU4HP5gPPvrRfgQ/puoL/qLpW+2sbQRoGfWyVxUN+Li3bT5BOMoaRai3vIX1/DqrIOgfYTNJopeawNUhBrZ5rHLak45DZzuaM+gy5KZSl5+naEVgv13y3y7W3ws8kfbb0fCvSBKJU7Kt8tv5h0txuuwPPkqbk2Rn4UsU673CaYwun2Qk2y+k15+lSus5nMvAD21NIn9uRpIklR5aDcz7Ze7c6Nj0XNeriQtu3kGqUh5I+06NL+c7IwQNqz+fV4bXnPNVOdIbEfF5RA6lfEQwqmfYqfHEQe6/ShIfFXEPVhjtuCZxNajJ5kY4HHpF+3L/LzQQ/I00x8g+lvoNv5h/1M3T8ob6XFFSOJp2JX0z7nEg/VZo7qJjK+wJJb5AuLLwG+ARp/qpaB8Azczvxu0hNb8V9G4YBp9u+uNp6JSeRftxvaum+yCE9l5bSbLR7kwL1Atqv5Wgh1XSaSE2TU23/UWlKkl+QmkK+43Rlf6fDJZWuEr+cVBv6Q66dCbhU0o9sX17Kuw1pSHRxElCep2tbUjNj0d7/WeBESQ/l92pv20+Xdt1KmjZlHPBoPjg3AWcpTb74W1JQ/6DtBcBxuXYxNq9fDgyV7115SvpaJzY/y8sm5xOGjSXdkLc7XNIOtouZsD9LuvDxC7T3jxwPjKa9JvQy6TO5nfR9+n1Ob6bj8bKp/Fypz2crUvNhuaxVv785z5GSdrJdeRIxaEQAqd/TpKkhytbP6fNJP5gfkuYuOt/2iUpzQ9WaC+pB0oHoV7arzbxb9u9KN+h5MPeXnABL2qnLV76vB0x3unp9BumAVti7dIHaicCJleP4leb3+mYn5fgr6WDwJeAbkl4gXSHd1cWSkKYN3yYHj9VJTWyb2f4yQ38urQvyXzPp+om2HDSKOc86NCHlYHwxgKRRSv1PVZuvcp6JpAk09yfVDC+UdFguy/7lA77SHGiHkM6uC0su9sz9e58gTffybVLteiSpT2UW+Qr5HAROJR1kF+RtziX9JpqBr7j9Atj/kjRe0pHARNIB+5Ok2l55doFHJH3U9pW5nP/I6Z1dJf9BYEFX1wcpTTuzme0DJZ0n6SN50XeAJZ+bbStNwz/MaX6vwkMsfZJ3vNK8ZyYdB6aSarXloNhMajUYR+qbfBvpPjECvm+71kW9g0IEkDo53dtgrqStbD+Qv5CvFD8SSd8knTn/g/Yr1ouroDtr66y3I+1A4PP5ANyStzuKNKV20RTza+AnSp3to0md/uT9l9uurXT3wUtt/6u0j6pzQZWcQDqAHEq6Wv4tpIDzEtBp/4fttxePlTogRzpNRghDfC6tHDD2BL4HzCzVJgxsZXslgBzofk46EM/Pfy+RvlMvVikHkt5KCoiH5YPoH/J7t3cuSzFIYE3Se/w8cIjzFexZh3mw8vfjItLn8i3Se7xmfv4+0o3FriFN8b6H7Q53UVS6EVSlS4Bfkmrdc0mfyTGk/rhzcp4TgUuU+u5mk/pNlipfmdM9Sk6U9GOnCSKXkmuSh9J+4vAV0iSixYCCyvuQvIs0xc5M0vvXQjron0Q6iSpez8XVApc6DrN/jTRH3hzgVVKz5T2k2mV5Pq//Js1C8BcGkQgg3XM4cJHSRHivUWq/dprB9vaK/MXBqrMD4DuUpqlYSHt1t5k099MVpXyLbH+4vKLS8NYl8xA5zYG0N0vrEECyTVj68/8/SlOoVDGBNFFgMXne85L+TmkKF6WO1O+Wmke65CE+l1bJGaWgWWz/9tL+nyN18C9F0sdIzY0d2H6ein6U4oCemzmLsj8P/HfRwV6h2nQvI4FHnG7ZDPB0/p5WDs/dj3SALJtCmjW4bEXgTrffAuDZvL0lw2fz9+rDLK2rk4iVK8ufa3gtthfmfpCi76IYfPGTXIMt1iuvP4z0Hf5xaXv7lbdfpQZaVp6f7nhSM1kHSsOCy69pPVIwHVQigHRD/vFVneCthsvp4gBoe6s6tzUst+0Woz+GkUaxPFzPulSfD+pXSrOmFsNlW0iBsdqPGNLZ6Jm5vbfoc7iPdFZZeJo0wqgm25fUUeahNpfWMOBrSiOjyp3w7y71/3RmId2fLFDk33g+4FULHlC9eexo0kF2FO01gL+TgmTZSjkQFB3xxcnKMbRPFU9e73ylOd/IeZ+m8xOWwlxSf18tTwE/Upoevo30mkeQbkPc2XVNv875j6Xj96AZ+LrSrRGKfsy3kT7betQ7OKk8Qm5QXhMSU5kMEkojslo9wKesVhoCeUEXZ2h9SgNgLi2lEUBNrphwUR1vZxsGgPx9oY6gvtyLABJCCKEhcR1ICCGEhkQACSGE0JAIICGEEBoSASSEEEJDIoCEEEJoSASQEEIIDfl/hJJnF+9q/+kAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def autocorr_resid(fitted_model):\n",
    "    from statsmodels.graphics.tsaplots import plot_acf\n",
    "    import matplotlib.pyplot as plt\n",
    "    from statsmodels.stats.stattools import durbin_watson\n",
    "\n",
    "    # 오차항의 자기상관성 검정\n",
    "    plot_acf(fitted_model.resid)\n",
    "    plt.xlabel('Lag at k')\n",
    "    plt.ylabel(\"lag at k's autocorrelation\")\n",
    "    plt.title(f'DW test 통계치 : {durbin_watson(resids=fitted_model.resid) : .5f}')\n",
    "    plt.text(y=-1.75,x=0, s='DW 통계치가 2에 근접하면 자기상관성이 없습니다.\\n0에 근접하면 양, 4에 근접하면 음의 자기상관성이 존재합니다.')\n",
    "    plt.show()\n",
    "\n",
    "autocorr_resid(fitted_model)   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2-3. 오차항의 정규성"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "검정통계치 :  8.35083\n",
      "p값 :  0.01537\n"
     ]
    }
   ],
   "source": [
    "# 정규성 검정 (귀무가설 : 정규성를 따른다)\n",
    "from scipy.stats import jarque_bera\n",
    "\n",
    "statistic, pvalue = jarque_bera(fitted_model.resid)\n",
    "print(f'검정통계치 : {statistic : .5f}\\np값 : {pvalue : .5f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jeong-giho/Desktop/financial_python/finance/lib/python3.9/site-packages/seaborn/distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 오차항 시각화\n",
    "sns.distplot(fitted_model.resid)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEICAYAAAB1f3LfAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAylElEQVR4nO3dd3hc1Zn48e+rMuq9WpZlWy5yAdvYwsaEYmODMVlCKKGFkGVDICGbhOQHm7AkC2lANiG7SSgJLNmQTSgpmBQwCQmmY4EwptlywZYtW7aq1bvm/f0x1zDI6tLozmjez/Poseaec++8xyqvzrnnniOqijHGGDNaEW4HYIwxJrRZIjHGGDMmlkiMMcaMiSUSY4wxY2KJxBhjzJhEuR3ARMvMzNQZM2a4HYYxxoSU119/vVZVs/orC7tEMmPGDEpLS90OwxhjQoqI7BuozIa2jDHGjIklEmOMMWMS8EQiIjkislFESkTkaRHJG6DeLSJSKiJbROTCfspXisgjA5w7V0RaRCR2vOM3xhgzuInokdwJ3KqqK4CbgB/2rSAi64EMVS0GTgauE5FMv/LHgIf7u7iIRAPfBXYGIHZjjDFDCGgiEZFUfAmiBEBVS4EU57i/q4E7nDodwP3AZUcLVfUCYNUAb/NN4KdA0yBxXOP0dkprampG1RZjjDH9C3SPZCawu8+xPc5xfzmqWun3ehcwe6iLi8ipQISqvjBYPVW9T1WLVbU4K6vf2WvGGGNGKdDTfwXob3nh4Sw5PGgdEUkBvgJcMoq4jDHGjJNA90jKObZnUegc91crIrl+r+dybE+mrzXAdOBpEXkWWAL8TUQ+OspYjTHGjEJAE4mq1gNtIrIUQEQWAXWAR0Tu9qv6IHCjU8cDXMkAN9f9rv2Yqi5T1VWqugrYCpylqk+Me0OMMcYMaCKebL8eeEBEEoFG4CogCV/PBABV3SAiy0WkBN+Q1m2qWjcBsRljBvFQyf6AXv/yFQUBvb6ZGAFPJKp6AFjX53AVsL5PvZuGuE45cOkg5atGF6ExxpixsCfbjTHGjIklEmOMMWNiicQYY8yYWCIxxhgzJpZIjDHGjIklEmOMMWNiicQYY8yYWCIxxhgzJpZIjDHGjIklEmOMMWNiicQYY8yYWCIxxhgzJpZIjDHGjIklEmOMMWNiicQYY8yYWCIxxhgzJpZIjDHGjIklEmOMMWMyEXu2G2NMv2xP+MnBeiTGGGPGxBKJMcaYMQl4IhGRHBHZKCIlIvK0iOQNUO8WESkVkS0icmE/5StF5JE+xxaIyO9EZJOIbBaRKwLVDmOMMf2biHskdwK3qmqJiBQDPwQu968gIuuBDFUtFpFY4AkReU5Va53yx4ClwOY+144HvqCq1SISA7wiIn9S1aZAN8oYY4xPQHskIpKKL0GUAKhqKZDiHPd3NXCHU6cDuB+47Gihql4ArOp7fVUtVdVq5/NOoBqI7ieOa5zeTmlNTc3YG2aMMeZ9gR7amgns7nNsj3PcX46qVvq93gXMHskbicj1wOuqWte3TFXvU9ViVS3OysoayWWNMcYMIdCJRADt53h/x0ZTBxFJFJGfA22qevNIgjPGGDN2gb5HUs6xPYtC57i/WhHJVdXDzuu5HNuTOYaIpAO/Bm5U1XfHFqoxxpjRCGiPRFXrgTYRWQogIouAOsAjInf7VX0QuNGp4wGuBB4exlvcDHzNkogxxrhnImZtXQ88ICKJQCNwFZCEr2cCgKpuEJHlIlKCb0jrtv7udfRjBbBMRPyPfUNVXxyv4I0xxgwu4IlEVQ8A6/ocrgLW96l30xDXKQcu7XPslHEI0RgzRocbOyjdV8++ujbqW7uIiBAyEz3MzkrkxJnpJMceM5nSTCK21pYxZtQa27v5y1uVvFvZRFSEMD0jnkX5KXhVOdTYwTNl1Ty3s4aPzM5kzbxsoiJtMY3JyBKJMWZUdlY188hr++npVdbMz2blzAziYz78K6WupZNNO3zJZGdVM586aTqp8R6XIjaBYn8eGGNGrGRvHQ++XE5avIcvr5nDmnk5xyQRgIzEGC5aNo1PnTSdI21d/Pz5PdQ0d7oQsQkkSyTGmBF5rbyeP26tZG5OEteeNouMxJghz5k/JZmrTymkx6v84qW9NLV3T0CkZqJYIjHGDFvZoSYef+Mgc3MS+eSKAjxRw/8Vkpcax1Unz6C9u5dfvVJOV483gJGaiWSJxBgzLFVNHTxSWkFeahyXL58+qhvnealxXHbiNA41dvCXtyqHPsGEBEskxpghdfV4eahkP57ICK44afqIeiJ9FeUmc/rcLEr3HeGtAw3jF6RxjSUSY8yQnnz7EDUtnVxcPI2UuLE/E7Jmfg7T0uL405uVtHb2jEOExk2WSIwxg9pZ1cyr5fWcOjuT2dmJ43LNyAjhgqX5dHT3svGdQ+NyTeMee47EmBD2UMn+gF6/q8fLH7ceJDMxhrULcsb12jnJsZw6J4vndtawbHo6MzMTxvX6ZuJYj8QYM6B/bK/iSFs3558wlegAPJW+uiiblLhonnz7EF4d1s4RJghZIjHG9KumuZOX3quleHpawHoLnqgIzpyfw8GGdt4+2BiQ9zCBZ4nEGNOvJ98+RHRkBGeO85BWX0sKUpmSEsvT26ro9VqvJBRZIjHGHGN3dQs7qpo5Y142SQFeuTdChLXzc6hv7eLNioaAvpcJDEskxpgPUVWe3naYlLhoVhZmTMh7zstNIi8llk07qq1XEoIskRhjPmRnVTMVR9pZVZQ1Ycu+iwhnzMumrrWLtw82TMh7mvFjicQY8z5V5e/bq0mLj2bZ9LQJfe95U5LJSorhxV21qM3gCimWSIwx79t+qJmDDe2cMS+bqIiJ/fUQIcIpszOpbOxgT23rhL63GRtLJMYYALyq/H17FRkJHpZMm9jeyFFLpqWSEBPFi7tqXXl/MzqWSIwxgG+J+MNNHZwxL5vICHElhujICFbMTGdnVTNHWrtcicGMnCUSYwwAL+yqJS0+mkX5qa7GUezcm3ltX72rcZjhC3giEZEcEdkoIiUi8rSI5A1Q7xYRKRWRLSJyYT/lK0XkkT7HEkTkURF5VUReEpEFgWqHMZNZRX0b++rbOHlWpmu9kaNS4z0U5SbxevkRmwocIiaiR3IncKuqrgBuAn7Yt4KIrAcyVLUYOBm4TkQy/cofAx7u59r/ATyqqsuBK4AfBSB+Yya9F3bXEhsd8X5vwG0nzkinubOHssNNbodihiGgiUREUvEliBIAVS0FUpzj/q4G7nDqdAD3A5cdLVTVC4BV/bzFaar6mFNnL1AuIsePbyuMmdzqW7t492Ajy2ekExMd6XY4AMzNSSIlLppX99rwVigIdI9kJrC7z7E9znF/Oarqv+/mLmD2YBcWkTSg79SOfs8TkWucYbPSmpqaYQVuTLh4+b1aRGDlrMyhK0+QyAhh2fQ0dle3UG833YNeoBOJAP0Ncg5n4HOoOsO+tqrep6rFqlqclZU1jLc2Jjy0d/VSWn6Exfmp47Lz4Xg6OsxWWm69kmAX6ERSzrE9hELnuL9aEcn1ez2XY3syH6Kq9UB2n8NDnmeM+cDr++rp6vXykdnB0xs56v2b7vvspnuwC2gicX7Zt4nIUgARWQTUAR4Ruduv6oPAjU4dD3Al/d9c76tERM51zpsK5KvqO+PYBGMmLa8qJXvrmZ4eT15qnNvh9GtpQRrNnT3sqW1xOxQziInYavd64AERSQQagauAJHw9EwBUdYOILBeREnxDU7epat0wrv0fzrVvAjqBa8c7eGMmqz01rdS1dnHGvL4d++BRlJtETFQEb1Y0Mic7ye1wzAACnkhU9QCwrs/hKmB9n3o3DXGdcuDSPscagYvGHqUx4adkbx3xnkiOm5ridigDio6M4Li8FN6pbOS8JXkB2e7XjJ19VYwJQ03t3Ww/1MSy6WlB/8t58bRUOnu8lB1udjsUM4Dg/g4yxgTEa/vq8Sosn5HudihDKsxKICkmynZPDGKWSIwJM71epbT8CHOyE8lIjHE7nCFFiLAoP4UdVc20d/W6HY7phyUSY8LMzqpmGtu7WTEz+HsjRy2elkqvV3m3stHtUEw/LJEYE2ZK9x0hKSaKotxkt0MZtqmpcWQkeNhqw1tByRKJMWGkpbOHHYebWFKQ6voqvyMhIiyelsre2laaO7rdDsf0YYnEmDCytaIBr/oe9As1x01NQYFth2xF4GBjicSYMKGqbNl3hPy0OHKSY90OZ8RykmLISPDwbqUlkmBjicSYMFHZ2MHhpg6WBcmeIyMlIhw3NYU9NS20dfW4HY7xY4nEmDCxZd8RoiKERVNT3Q5l1BbmJeNVKDtkDycGE0skxoSBnl4vWysaWJCXTJwnODavGo2pqXGkxEXbNOAgY4nEmDCw/XAz7d29IXmT3Z+IsDAvmV3VLXR228OJwcISiTFhYMu+IyTHRjE7O9HtUMZsYV4KPV5lR5UNbwULSyTGTHJNHd3srGrmhII0IiR0nh0ZyPSMeBJiomz2VhCxRGLMJLd1fwMKLAvxYa2jIkRYOCWZHYeb6e71uh2OwRKJMZOaqrJl/xEK0uPJTAr+BRqHa0FeMl29XvbWtrodisESiTGTWmVjB9XNnSF/k72vmZkJeCIj2G5PuQcFSyTGTGJvVjQQKcJxU0NngcbhiI6MYHZ2IjsON6OqbocT9iyRGDNJeVV580ADc3OTiPcEfFftCTcvN4mG9m4ON3W4HUrYs0RizCS1p6aV5o4elkxLdTuUgJibmwTADtuC13WWSIyZpN6saCAmKoJ5zi/cySY5Npr8tDi7TxIEAp5IRCRHRDaKSImIPC0ieQPUu0VESkVki4hc6Hc8UkTuEZHXnI9T/cpSReT3IrLJOfdzgW6PMaGgu9fLO5WNLMxLITpy8v69WJSbxIEj7bR02iKObhr2d5iIjHbu4J3Araq6ArgJ+GE/114PZKhqMXAycJ2IZDrF1wC7VfVEYB3wXRGJdsq+DjymqquBU4HP+51nTNgqO9xMZ4930g5rHTU/NxnFhrfcNpI/VTaKyK9F5OPDTSoikoovQZQAqGopkOIc93c1cIdTpwO4H7jMKbsEuNspqwc2Amc7ZZHAAefzDqACaBtBm4yZlLZWNJAUG0VhVoLboQTUlJRYkmOjKDtsw1tuGvZUDlU9wxmW+hjwaxFpAf4EPKmqnQOcNhPY3efYHuf4G37HclS10u/1LmDl0Rj7XH8XMNv5/HbgHhE5CTgJ+IGqHpNIROQafD0bCgoKBm+oMSGurauHnYebOakwfVIsiTIYEaEoN5k3DzTQ0+slahIP4wWzEf2vq2qlqv4M3/CUAF8EfisiXx3gFAH6m+Q9nInfg9U5WnYR8DrwM3zDZv8qIov6ifs+VS1W1eKsrKxhvLUxoeudg030qrJk2uR6CHEg83OT6OrxsrfOnnJ3y0jukcwQkW+KyNPAOcC3VfUMVT0PuHSA08r5oPdwVKFz3F+tiOT6vZ7LBz0Zr4h4+paJSCJwkar+QFWbVHUH8CPgquG2yZjJaGtFA5mJMeSlht52uqNRmJVIVIRQZvdJXDOSHsn3gFeAs1T1FlXd41f2z/2d4NzTaBORpQBOb6EO8IjI3X5VHwRudOp4gCuBh52yDcC1TlkysBb4K9AFTD2agEQkAvgkHx4yMyasNLR1UV7XypJpKcgkH9Y6yhPle8q97FCTPeXukpE87rpdVf/uf0BEblLV21V12yDnXQ884PQgGvH1GJLw9UwAUNUNIrJcRErwDVvdpqp1TvFdwL0ishnoBW5Q1W7n/T8HPCIiiu/G+x9V9VcjaJMxk8qbB3w7By7OT3U3kAlWlJtE2eFmqps7yUkOj55YMBkykYhIAb6b4xeIyPN+RTH4egC3D3a+qh7AN23XXxWwvk+9mwY4vxvfrK7+yl4AVg32/saEkzcrGihIjycjcfKs9Dsc83KT+SOVlB1utkTiguH0SGbi+2Wd5vx7tL/sBewBQGOCxOHGDg43dXDu4n6f+Z3UUuKimZISS9nhJk6faxNqJtpwEslLqvqciGzBd2/CGBOEtlY0ECFw/NQUt0NxRVFuEs/vrKG9q5c4T6Tb4YSV4dxsv9n596vAU/18GGNc5lXlrQMNzMlOIjFm8q30OxxFOUl4FXZV2+ytiTbkd5yqfsv59DwgUlWPiMiJQDHwv4EMzhgzPPvq2mho7+ashblDV56k8tPiiYuOZGdVM4vCbLKB20Yy/fchIFtE0oH/xHfPxBKJMUFga0UDnsgIFkyZXBtYjURkhDAnJ5GdVS14bRrwhBpJIkl3Hvo7G7hHVW8DpgQmLGPMcPX0ennnYCML8pLxRIX3EiFFOUm0dPZwqME2u5pII/quE5FC4LN8cG/E0r4xLttZ1Ux7d2/YPTvSnzk5SQiwo8oWcZxII0kkX8G3JPx9qtosIqcB/xeYsIwxw7W1ooEETySzsxPdDsV1iTFRTE2Ls2XlJ9iwE4mzFPx9QKKzmu48RvZkvDFmnHV091J2uJnj81OJjAiPJVGGUpTj2+yq1Ta7mjAjWbTxfuBcfGtcdfp9GGNc8m5lIz1e5YRJvoHVSBTlJqHYNOCJNJIexXxVPSVgkRhjRmxrRQPpCR7y0+LcDiVo5KXGkeCJtOGtCTSSeyT1IhIeGxwYEwKqmjrYU9PKkmmpYbPS73BEiDA3J4mdVS30em0+0EQYSY9kE/C8iDwIVB89aKvtGuOOP22tRIElNlvrGHNzk3ijooE3DzSwtMD+/g20kSSSLc6H8sHCjcYYlzy+9SBTU+PITAqvlX6HY052IgI8W1ZtiWQCjGRoaxu+nRD/TVWfw7cqsCUUY1ywu7qZdyubWGI32fsV74miID2eTTtq3A4lLIwkkTwA/AaId14/hG+fdGPMBHv8jUoiBBblh+dKv8NRlJvE2wcbqW62p9wDbSSJJFZVX8TphahqF+AZ/BRjzHhTVR7fepCPzM4kKTba7XCC1tycJACes15JwI0kkVQ5T7MrgIisB/YFJCpjzIBe33eEA0fa+fiSqW6HEtSmpMSSnRTDs5ZIAm4kieQHwDeAIhF5CfgY8K8BicoYM6DHtx4kNjqCdceF75LxwyEirCrK4vldNXT3et0OZ1IbMpGIyOdFZBvwHaAWKAeygc2q2hLY8Iwx/rp7vTzx1iHWzs8J2w2sRmJ1UTbNHT1s2XfE7VAmtUG/E0XkK8AKYLl/0hCRfOB+EelQ1UcDHKMxxvHsjhqOtHXbsNYwfWROJlERwrM7a1hRmOF2OJPWUD2SS4Er+/Y8VPUAcDFw3VBvICI5IrJRREpE5GkRyRug3i0iUioiW0TkQr/jkSJyj4i85nyc2ue8r4rISyKySUR+PFQ8xoSy379eQWaih9OLstwOJSQkx0ZTPCONTWXVQ1c2ozZUIml3ZmcdQ1WbgeEsr3kncKuqrsA3XfiHfSs4N+4zVLUYOBm4TkQyneJrgN2qeiKwDviuiEQ7510OFAKnqepq4L+HEY8xIamupZN/bK/m/BOmEh0Z3htYjcTqomzKDjdzqLHd7VAmraG+G9NE5KwBPtYBg/YVRSQVX4IoAVDVUiDFOe7vauAOp04HcD9wmVN2CXC3U1YPbMS3SyPAl4CvqWqvU753qAYbE6r+uLWSHq9y0bJpbocSUlbPywaw2VsBNNTduseAlYOUbxji/JnA7j7H9jjH3/A7lqOqlX6vd/m9b5SqdvYpmy0isUALcK6IfAaIAe5V1Yf7BuHsn3INQEFBwRAhGxOcfvf6ARblp1CUm+R2KCFlTnYiU1Pj2FRWzWXL7ec/EAZNJKr6rTFeX+h/O97hLMk5WB0FEoDFQDq+Ia8E4AkR2eLsLf9BZdX78G3KRXFxsS0HakLOu5WNbD/UxLfPW+h2KCHn6DTgx984SGdPLzFRkW6HNOkEeqC1HJjd51ihc9xfrYj4T4qfywc9Ga+IePqWqWodsFdV71FVr3PP5k/A8eMVvDHB4nelB/BERvCxxf3OVTFDWF2UTWtXL6XlNg04EAKaSJx7Gm0ishRARBYBdYBHRO72q/ogcKNTxwNcCRwdotoAXOuUJQNrgb86ZX9zhq1wbsCvwbdCsTGTRlePlz9uPciZC3JIjbdViUbj5NkZeCIjbPZWgEzEE03XAw+ISCLQCFwFJOHrmQCgqhtEZLmIlOAbtrrN6XEA3AXcKyKbgV7gBlXtdsq+A9wpIs/jG0a7S1X3TECbjJkwz5RVcaStm4uK890OJWTFe6JYUZjOsztr+IbbwUxCAU8kzjMn6/ocrgLW96nX70rCTtK4eoCyTmyZFjPJ/a70ANlJMZw6O3PoymZAq4qy+c5ftlFR38a09PihTzDDZpPRjQli1c0dPLuzhguW5hNlz46MyWrnIc5nd9jw1niz70xjgtjjbxyk16tctMyGtcZqZmYC0zNss6tAsERiTJBSVX5beoATClKZnZ3odjghT0RYXZTNy+/V0tHd63Y4k4olEmOC1GvlR9hd3cJlJ9pDdONlVVEWHd1eNu+pG7qyGTZLJMYEqd+U7CMpNop/WjzF7VAmjZMKM4iNjrDlUsaZJRJjglBdSycb3z7MhUvziffYviPjJTY6kpWFGTxTVo2qLXIxXiyRGBOEfv/6Abp6vVy+woa1xtvqednsr29jb22r26FMGpZIjAkyXq/y0Kv7WT4jnbk5tkDjeFs111YDHm+WSIwJMi+9V8u+ujY+eZL1RgKhICOeWVkJbLLnScaNJRJjgsxvNu8nPcHD2cflDl3ZjMrqomxK9tTT1jWcvfnMUCyRGBNEDhxp42/bDvOJ4nxb7jyAVs/LpqvXy8u7bRrweLBEYkwQ+dUr+xARPr1yhtuhTGrFM9JI8ETa8NY4sURiTJBo7ezh4Vf3s/64XPJS49wOZ1KLiYrkI7Mz2WTTgMeFJRJjgsTvXz9Ac0cP/3LKTLdDCQtr5+dQ2djBu5VNbocS8iyRGBMEvF7lf1/aywkFqSwtSHM7nLCwZn42EQJ/21bldighzxKJMUHgmbJqyuva+Iz1RiZMRmIMxdPT+du7h90OJeRZIjEmCDzw4l7yUmI5e6FN+Z1IZy3MoexwMxX1bW6HEtIskRjjsi37j/DKnjqu+shM27xqgp25IAew4a2xsu9aY1x21zO7SYuPtnW1XDA9I4GinCQb3hojSyTGuOidg408U1bNZ06ZSUKMrfLrhrMW5vBaeT31rV1uhxKyLJEY46K7N+0mKTaKK0+e4XYoYevMBTl41TfhwYyOJRJjXLKzqpmN7xzmqpNnkBwb7XY4Yev4qSnkJsfa8NYYBDyRiEiOiGwUkRIReVpE8gaod4uIlIrIFhG50O94pIjcIyKvOR+n9nPuXBFpEZHYQLbFmPF096bdJHgiueojNuXXTSLCWQtzeH5XDa2dtojjaEzEoOydwK2qWiIixcAPgcv9K4jIeiBDVYudZPCEiDynqrXANcBuVb1ORNKBDSKyVlW7nXOjge8COyegLaaPh0r2B/T6k/UGdNnhJv78ZiWfPa2QtASP2+GEvfXHTeFXr+xj045q/mlRv3/rmkEEtEciIqn4EkQJgKqWAinOcX9XA3c4dTqA+4HLnLJLgLudsnpgI3C237nfBH4KDLjOgYhc4/R2SmtqbDMb477vbywjMSaKz58+y+1QDLB8ZjqZiTE8+fYht0MJSYEe2poJ7O5zbI9z3F+Oqlb6vd4FzHY+j1LVzv7KnGGuCFV9YbAgVPU+VS1W1eKsrKyRtsGYcfXye7Vs2lHDF1bPJjXeeiPBIDJCWH9cLs+UVdseJaMQ6EQiQH9Law5nuc3B6qiIpABfAb41msCMcYPXq9yxsYypqXF82mZqBZWPLppCR7fXZm+NQqATSTkf9CyOKnSO+6sVEf+1IebyQU/GKyKefsrWANOBp0XkWWAJ8DcR+eh4BG5MIPzl7UO8daCR/3fWXGKjbeOqYHLiDBveGq2AJhLnnkabiCwFEJFFQB3gEZG7/ao+CNzo1PEAVwIPO2UbgGudsmRgLfBXVX1MVZep6ipVXQVsBc5S1ScC2SZjRquju5cf/LWM+VOS+fiSqW6HY/qIjBDOOd6Gt0ZjIp4juR64XURewndD/UYgCV/PBABV3QB0iUgJ8Dxwr6oe3QPzLmCxiGzGd6P9G0dnbBkTSu7etJuK+na++dH5RESI2+GYfpxzvA1vjUbAp/+q6gFgXZ/DVcD6PvVuGuD8bnyzuoZ6n1WjDNGYgNtV1czPnnuPC06YysmzM90OxwzAf3jLpgEPnz3ZbkyAeb3Kv294m4SYKG7+6Hy3wzGD8B/earGHE4fNEokxAfZoaQWvlR/h38+ZT0ZijNvhmCGctySPjm4vT71jS6YMlyUSYwLoUGM7tz+5nZMK0/nEsny3wzHDsLQgjekZ8Wx444DboYQMSyTGBEhPr5cvP7yVHq9y+wWLELEb7KFARPj4kqm8/F4dhxrb3Q4nJFgiMSZAfvrMbl4tr+d75x/HzMwEt8MxI3D+CVNRhT9urRy6srFEYkwgvPJeHT99ZhcXLs3n/BNsSCvUzMhMYGlBKhu2HER1OAtxhDdLJMaMs+rmDq5/9A1mZCbw7fMWuh2OGaXzl+azo6qZbYcGXA/WOCyRGDOOOrp7+eyvXqepvYe7Lltq2+eGsH86fgrRkcKGLQfdDiXoWSIxZpz0epWvPLqVtw408N+XLmFBXrLbIZkxSEvwsLoom8e3VtLT63U7nKBmicSYcaCq3LzhbTa+c5ibz5nPuoW5Q59kgt6Fy/Kpbem0JVOGYInEmDHyepVvPP4Oj7xWwRdWz+LqUwuHPsmEhDXzsslJjuGhVwO7E2ios0RizBh09Xi54Xdv8puS/Xx+1SxuOKvI7ZDMOIqKjOCS4mk8t7OGivo2t8MJWnYn0IS1sew539bVw0Ov7mdPTStr5+eQnxrHw69WjGN0JhhcsryAuzbt5tHXKrhhnf2h0B/rkRgzChX1bdy1aTf7atv4xLJ8zpiXbU+uT1JTU+NYVZTNo6UVdNtN935ZIjFmBHp6vTy9rYqfP/8eKFx7eiEnFKS5HZYJsMuXF1DT3Mnft1W5HUpQsqEtY4ZBVdl+qIkn3zlMfWsXJ0xL5Z8W5RHnse1yw8GqoiympMTy0Kv7WX/8FLfDCTqWSIwZRHevl3cONvLye3UcbGgnKzGGq06ewZycJLdDMxMoKjKCS08s4L/+vpM9NS0UZiW6HVJQsURiXNPe1Ut9axf1bV00tHXR3t1LZ7eXzp5eVEEE3th/hISYKJLjokmJiyY5NorUeA+ZiR6ykmLITIwhNnp8ewVdPV7217ex7VAjWysa6Oj2kpHg4cKlU1kyLY1I2yY3LF2+ooC7N+3mgRf38r3zj3c7nKBiicRMCFWlprmTXdUtVBxp48CRdupbuz5UR4DY6Eg8URFECKhCZUMHrV09NHcMvFtdUmwUWYkxZCbFkJUUQ1ZijJNkPkg2WUkxJMVGEymCCPR4leaObqqaOmju6KG2pZO6lk4qGzvYX99Gr1eJihCOm5rCsulpzMxMIMJupoe1rKQYLlg6ld+/foCvnjnXNinzY4nEBIyqUtnQwdaKI2w/3Px+4kiNiyY/LY7lM9LJTPSQluAhNc5DbHTEMTOfLl9RAPiWH2np6KGpo5sjbV3UtnRS09xJbUsXNc2+z2taOtle2cTzLZ2DJp6BREcK2UmxnFyYQWFWIjMy4okZ596OCW1XnzqTR16r4P827+P6tXPdDidoWCIx466zu5ctFQ2UltdzqLGDqAihMCuBU2ZnMi83idR4z4ivGRkhpMRHkxIfzbT0+CHrd3T3vp9sjiacls5uvApeVSJFSIqN5q0DDSTERJGZGENSbJT1OsygZmcnsXZ+Nr96ZR/XnjbLJls4Ap5IRCQH+CWQDjQBn1bVY3aLEZFbgHPxTUn+nqr+wTkeCfwUONGp+lVVfcEpWwB8C8gE4oC7VPXXAW2QGVBbVw8vv1fHK+/V0d7dS15KLB9bnMfi/NQJ/4GLjY4kPy2e/LShk44xI/HZUwu55L7N/GHLAa44abrb4QSFieiR3AncqqolIlIM/BC43L+CiKwHMlS1WERigSdE5DlVrQWuAXar6nUikg5sEJG1qtoNxANfUNVqEYkBXhGRP6mqbSAwgTp7enl+Zy0v7a6lq9fL/CnJrJqbNayegzGhZvnMdBZPS+WBF/dy2fICm3xBgB9IFJFUfAmiBEBVS4EU57i/q4E7nDodwP3AZU7ZJcDdTlk9sBE4++j1VLXa+bwTqAaiA9ci46/Xq5SW1/Ojv+1k045q5uYm8aU1c/jUSdMtiZhJS0S49rRC9ta28pe3bCteCPyT7TOB3X2O7XGO+8vpM9y1C5jtfB7lJIn+yt4nItcDr6tqXT9l14hIqYiU1tTUjLAJpj/vVjZywT0v8dgbB0mNj+ZzpxVy+fICcpNj3Q7NmIA7e2Eu83KT+K+nd9qyKQQ+kQjQ34bHw9kEebA675eJSKKI/BxoU9Wb+62sep+qFqtqcVZW1jDe2gykvauX2zdu52N3vcTBhnYuLp7G506fRUFGgtuhGTNhIiKEG84qoryujT+8fsDtcFwX6ERSzrG9h0LnuL9aEfHfCWguH/RkvCLi6a/MuWfyW+AnqnrfOMVsBlB2uIlz73qRnz+3h4uW5vP3r57OkmmptlihCUtr5mezZFoqP/7HLjq6e90Ox1UBTSTOPY02EVkKICKLgDrAIyJ3+1V9ELjRqeMBrgQedso2ANc6ZcnAWuCvTtnNwNdU9d1AtiPcqSq/KdnHeXe9RGN7N7/+zAq+f9GiUU3jNWayEBH+bV0Rhxo7+M0YtiOYDCZi1tb1wAMikgg0AlcBSfh6JgCo6gYRWS4iJfiGrW7zu9dxF3CviGwGeoEbnBlbACuAZX3+Iv6Gqr4YyAaFk8b2bv79sbd54u1DnDonkx9dvISsJHui1xiAk2dncvKsDO7ZtJtLTpxGYkx4PpoX8Far6gFgXZ/DVcD6PvVuGuD8bnyzuvorO2U8YjT921rRwL8+tIVDjR187ex5XHtaIRE21dGYD7lxXRHn3/MyP31mFzetn+92OK6w/UjMMbxe5b7n3+Oie19GFX577Uo+v2qWJRFj+nFCQRoXF+fzwAt72VnV7HY4rrBEYj6krqWTf3nwNW57soy183N48kunsmy6bdxkzGC+vn4+ibFRfOPxd1AdzqTUycUSiXnfK+/Vcc5PXuDl9+r4znkLufeKpaTE2/OdxgwlPcHD18+ex6t763lsy0G3w5lw4XlnyHxIr1f58T928dNndjEzI4Ff/POJLMxLcTssAB4K89kwJnRcXDyN35ZWcNuT21kzPzusZjVajyTMHWps57L7N/OTf+zi/BOm8ucvnhI0ScSYUBIRIXz348fT2N7NzRvCa4jLEkkY+8f2Ks758Qu8c7CROz+xmB9dvISEMJ2+aMx4WJCXzA3rinji7UM8+lqF2+FMGPutEYa6erx8/6kyHnhxL/OnJHPX5Scwy/agNmZcXHNqIS/uquVbf95G8Yw0ZmcnuR1SwFmPJMzsqWnhop+9zAMv7uXKldPZcN3JlkSMGUcREcKPLl5MnCeSLz68NSyWT7FEEiZUlUde3c9Hf/Ii++ra+NkVS/n2eccRa1vJGjPuspNjufMTi9l+qCks7pfY0FYYONLaxdcfe4u/vlvFR2ZncOcnlpCbYsu9GxNIq+dl8+U1c/jxP3YxKzuB61Yds/vFpGGJZJLbVFbN1x97i/rWLm4+Zz6fOWWmPaFuzAT58po5vFfTwn8+tYOUuGg+uWJybs1riWSSqm3p5Nt/3saf3qxkdnYiD3z6RI6batN6jZlIERHCnRcvprWzh5s3vIMnMoJPFE9zO6xxZ4lkklFV/rDlIN99YhutnT1cv3YOn181i5gouxdijBtioiK594plfPZXpfzbH97CExXBeUumuh3WuLJEMonsrGrm23/exou7a1k2PY07LjieOTmTf+qhMcEuNjqS+z5VzFW/fJXrH91KZUMHnzu9cNJsCmeJZBKobenkv57eycOv7icxJorvnLeQT66YbvdCjAkicZ5IfnnVcm78/Vt8/6kydlY1c/sFx0+KmZOWSEJYQ1sX//PCXn75cjnt3b1cuXIGX14zh7SE8Fnjx5hQEhsdyU8uXcLc7ETufHone2pauPPiJczODu1nuSyRhKDqpg4efKWcB1/eR0tnDx89fgpfOXNuyH8zGhMORIQvrpnDnJwkvvaHtzjnJy/wlbVz+eypM4mKDM1H+yyRhAhV5c0Djfzypb088fYherzKOcdP4UtnzKEo1+6DGBNqzj4ul6XTU/nm4+/w/afKeOLtSv5t3TxOnZMZcvdOLJEEucONHTy+9SCPbTnAzqoWEmOiuOKk6Xx65QxmZCa4HZ4xZgyyk2L52RXLePLtw3zviW1c+YtXOXFGGl9ZO5eVszJCJqFYIgkyqsqu6hb+vr2Kv2+r4o2KBlRhaUEq3/n4cXx8SR5JsbbZlDGThYjw0UVTWLsgm9++VsFdm3Zz+f+UMCsrgUtPLOD8pVPJTIxxO8xByWRfA6av4uJiLS0tdTuM93V097KrqoU3DzRQsreezXvqqGnuBGBRfgpr5+dw7uI8ZgZp78M2njLB7PIVBW6HMGId3b386c1KfvtaBaX7jhAZISybnsYZ87JZVZTF3OwkV2ZkisjrqlrcX1nAeyQikgP8EkgHmoBPq2plP/VuAc7Ft5Dk91T1D87xSOCnwIlO1a+q6gtOWQLwC2Am0A18VlW3BbRBo6CqHGnr5sCRNirq29lf38bOqma2VTaxu6aFXq8vmeckx7CyMIOTCjNYPS+LKSlxLkdujJlosdGRXFw8jYuLp7G7upnH36jkH2XV3LGxjDs2lpEUG8WSaakszk9lbm4Ss7ISKMxMJM7j3jTiiRjauhO4VVVLRKQY+CFwuX8FEVkPZKhqsYjEAk+IyHOqWgtcA+xW1etEJB3YICJrVbUb+A/gUVV9TERmAvcCZweiEe1dvdQ0d9Le3Ut7dy8dR//tOvraS3NHN/VtXRxp7aK+tZsjzufVzZ20dPZ86Hq5ybEsyEvmzAU5LMhLZmFeMgXp8SEzJmqMCbzZ2UncsK6IG9YVcaixnRd21fLG/ga2VjRwz7O78foNKKUneMhJjiUnOYacpFiykmJIio0iKTba+TeKwsxECjLixz3OgCYSEUnFlyBKAFS1VERSRCRVVRv8ql4NfNGp0yEi9wOX4euJXAKsc8rqRWQjvmTxZ+A0Vf2aU7ZXRMpF5HhVfXu82/L37VV88eE3hqzniYwgPcFDWoKH9IRoFuQlc1piDNPS48lPi2NaWjz56XEk230OY8wITEmJe7+nAr4hsPK6Vt6rbmVPTQuHmjqobuqgqqmTbZVN1LV2vT/acdTnTp/F19fPG/fYAt0jmQns7nNsj3Pc/7dyTp/hrl3ASufzKFXt7FM2W0TSgNo+194FzAY+lEhE5Bp8PRuAFhHZMdKGjEBmP3FNRtbOyWPStvGTH345advpZ9A23vR9uGn01x5w6eJAJxIB+rubP5w7/IPV0ZFcW1XvA+4bxnuOmYiUDnRDajKxdk4e4dBGCI92utXGQD9GWY6vh+Cv0Dnur1ZEcv1ez+WDnoxXRDx9y1S1Hsjucx3/84wxxkyAgCYS55d9m4gsBRCRRUAd4BGRu/2qPgjc6NTxAFcCDztlG4BrnbJkYC3wV6esRETOdcqmAvmq+k4g22SMMebDJmLW1vXAAyKSCDQCVwFJ+HomAKjqBhFZLiIl+IamblPVOqf4LuBeEdkM9AI3ODO2wDdr6wERuQnoxEk4LpuQIbQgYO2cPMKhjRAe7XSljWH3QKIxxpjxFZpLTRpjjAkalkiMMcaMiSUSY4wxY2KJZAxE5CYR+Zzf63wReVdEnvX7ONOvfKWIlIjIayJyv4iExOrLfdvpHCsSkedE5FUR+YMzmeJoWUi2E0BEdvT5+n3Kr2zANocqEckRkY3O1+tpEclzO6axCoevYT+/e9z9eVRV+xjhBzAV2AzUAJ/zOz4DeGSAc6KB5/EtGQPwZeALbrdlNO10yp4CCp3PzwN+EKrt7NOuzYOU9dvmUP4Afg2scD4vBh5yOyb7Gg7atoF+97j682g9klFQ1YOqehLOsy/DdDbwlH4wrfle4KJxD24cDdROETke2K+qe5x6fwROdopDrp3DMUSbQ1J/a+EBKc7xSWcyfA37+5kMhp9HSyTjb5aI/FpENonIL0Uk0zk+G3h/jS9V7cL310Io+lBbHPXO+meh3s5MEXlARP4hIr8XkRnO8cHaHKoGWwsvlIXT1xCC4OcxZMauQ0QtvmXz/6yqrSJyAfBjfGvH9bc2WKg+xDPYOmeh3s5vAU+ob6Xpk4D/BVYztnXjgtVkbBOE19cQguDn0XokQxCRr/rdtPvWYHVVtUVVH1HVVuf1Y8Asp3gPvrXAjl7XA/QcexV3jKSd9GmLI0N9WwMEdTv99ddmVf0/9S3tg6puBo7uLjZYm0NVOcNbCy+khNnXEILg59ESyRBU9Uequsr5uGWwuiKSLCIL/V5fDrzivHwKONevG/0vwG8DEvQojKSdqroVmCsi0wFE5GzgBac4qNvpr782i8j7Y+YicjpQ4dTdysBtDkk6wFp4of6LNZy+hhAcP482tDW+eoH/J74FJD3ANuAGeH/DrpuBp0SkB3gTZzOvEPVl4NciIsBBfN+gk6Gdl4hv2+dooBLwn/bcb5tD3PUcuxZeqAu3ryG4/PNoa20ZY4wZExvaMsYYMyaWSIwxxoyJJRJjjDFjYonEGGPMmFgiMcYYMyaWSIwJMiISKyLPjuK8zQEIx5ghWSIxxkUicrmz6J4xIcseSDTGRar6kNsxGDNWlkiMGScicisQj2+BwP/Gt1x3MlANfMYp+z8gFd9iep/Ht3fEZuAt4H6nzkHnev8MxKrqz5zXd+Bb8qIRuA1IxLcu1j+ram+g22fMQGxoy5jxtRxYCVwK/LuqrgGeBL4EXAH8QVVXAuuBTr/z/ge4RVVXA7cP8R6twMdV9VSgGVg3vk0wZmSsR2LM+PoFEItvY6F7fUsfEQmU4eul/FJEkoA/qeo2v3KPs7EUwHuDvYGq7hSRNSJyGrAY3/pJxrjGEokx46sa334PW1T1zL6Fzsq0pwK3+83MisK34OdRKc6/R/eTOCrTucZ/4Rv+ut95P/86xkw4G9oyZpw5+9F0isgSABFJcWZnnYiv5/EMvg3PVjmndAJxIrLAef0Z599yfENliMhM4ALn+EnAPfj27b4YSyTGZdYjMSYwPgvcLSLp+O5p3AwkAI+JyNGtTj8HXO5X/+ciEgE8DnSr6nMicoWIvAzUA7906v4AeA7f/ZHfAV8RkaDc88WEB1tG3hhjzJjY0JYxxpgxsURijDFmTCyRGGOMGRNLJMYYY8bEEokxxpgxsURijDFmTCyRGGOMGZP/D4ehfxKUg/x4AAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "검정통계치 :  8.35083\n",
      "p값 :  0.01537\n",
      "오차항이 정규분포를 따른다는 귀무가설을 유의수준 5.0%에서 기각합니다.\n"
     ]
    }
   ],
   "source": [
    "def normal_test(fitted_model, sig_level):\n",
    "    # 정규성 검정 (귀무가설 : 정규성를 따른다)\n",
    "    from scipy.stats import jarque_bera\n",
    "    import seaborn as sns\n",
    "    import matplotlib.pyplot as plt\n",
    "    import warnings\n",
    "    warnings.filterwarnings(\"ignore\")\n",
    "    \n",
    "    sns.distplot(fitted_model.resid)\n",
    "    plt.xlabel('residual')\n",
    "    plt.show()\n",
    "    \n",
    "    statistic, pvalue = jarque_bera(fitted_model.resid)\n",
    "    print(f'검정통계치 : {statistic : .5f}\\np값 : {pvalue : .5f}')\n",
    "    if pvalue > sig_level:\n",
    "        print(f'오차항이 정규분포를 따른다는 귀무가설을 유의수준 {sig_level * 100}%에서 기각하지 못합니다.')\n",
    "    else:\n",
    "        print(f'오차항이 정규분포를 따른다는 귀무가설을 유의수준 {sig_level * 100}%에서 기각합니다.')\n",
    "        \n",
    "normal_test(fitted_model, .05)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. 변수선택법"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[*********************100%***********************]  5 of 5 completed\n"
     ]
    }
   ],
   "source": [
    "# 삼성전자, 하이닉스, 네이버, 카카오 KOSPI 데이터 불러오기\n",
    "stock_price = yf.download({'005930.KS','000660.KS','035420.KS','035720.KS','^KS11'}, '2021-12-01', '2022-01-21')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3-1. 전진선택법"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 하나씩 해보기\n",
    "## 전진 단계별 선택법\n",
    "\n",
    "import statsmodels.api as sm\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "# 변수 설정\n",
    "target = stock_price['Adj Close']['^KS11']\n",
    "ks11 = []\n",
    "ks11.append([stock_price.columns[5*i+4] for i in range(6)])\n",
    "feature = stock_price.drop(columns=ks11[0])\n",
    "variables = feature.columns.tolist()\n",
    "y = target\n",
    "\n",
    "# 선택된 변수들 list 생성\n",
    "forward_valriables = []\n",
    "\n",
    "# 전진선택시 P 값을 고려할 때, 선택과 제거 임계치 설정    \n",
    "sl_enter = 0.05\n",
    "sl_remove = 0.05\n",
    "\n",
    "# 각 스텝별로 선택된 변수들\n",
    "sv_per_step = [] \n",
    "# 각 스텝별 수정된 결정계수\n",
    "adj_r_squared_list = []\n",
    "# 스텝\n",
    "steps = []\n",
    "step = 0\n",
    "\n",
    "while len(variables) > 0:\n",
    "    remainder = list(set(variables) - set(forward_valriables))\n",
    "    pval = pd.Series(index=remainder) ## 변수의 p-value\n",
    "    ## 기존에 포함된 변수와 새로운 변수 하나씩 돌아가면서 \n",
    "    ## 선형 모형을 적합한다.\n",
    "    for col in remainder: \n",
    "        X = feature[forward_valriables+[col]]\n",
    "        X = sm.add_constant(X)\n",
    "        model = sm.OLS(y,X).fit(disp=0)\n",
    "        pval[col] = model.pvalues[col]\n",
    " \n",
    "    min_pval = pval.min()\n",
    "    if min_pval < sl_enter: ## 최소 p-value 값이 기준 값보다 작으면 포함\n",
    "        forward_valriables.append(pval.idxmin())\n",
    "        ## 선택된 변수들에대해서\n",
    "        ## 어떤 변수를 제거할지 고른다.\n",
    "        while len(forward_valriables) > 0:\n",
    "            selected_X = feature[forward_valriables]\n",
    "            selected_X = sm.add_constant(selected_X)\n",
    "            selected_pval = sm.OLS(y,selected_X).fit(disp=0).pvalues[1:] ## 절편항의 p-value는 뺀다\n",
    "            max_pval = selected_pval.max()\n",
    "            if max_pval >= sl_remove: ## 최대 p-value값이 기준값보다 크거나 같으면 제외\n",
    "                remove_variable = selected_pval.idxmax()\n",
    "                forward_valriables.remove(remove_variable)\n",
    "            else:\n",
    "                break\n",
    "        \n",
    "        step += 1\n",
    "        steps.append(step)\n",
    "        adj_r_squared = sm.OLS(y,sm.add_constant(feature[forward_valriables])).fit(disp=0).rsquared_adj\n",
    "        adj_r_squared_list.append(adj_r_squared)\n",
    "        sv_per_step.append(forward_valriables.copy())\n",
    "    else:\n",
    "        break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>          <td>^KS11</td>      <th>  R-squared:         </th> <td>   0.887</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.876</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   81.12</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Wed, 30 Mar 2022</td> <th>  Prob (F-statistic):</th> <td>9.03e-15</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>16:51:29</td>     <th>  Log-Likelihood:    </th> <td> -146.38</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td>    35</td>      <th>  AIC:               </th> <td>   300.8</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td>    31</td>      <th>  BIC:               </th> <td>   307.0</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>     3</td>      <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "               <td></td>                 <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th>                      <td>  734.6837</td> <td>  187.354</td> <td>    3.921</td> <td> 0.000</td> <td>  352.572</td> <td> 1116.795</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>('Low', '035420.KS')</th>       <td>    0.0020</td> <td>    0.000</td> <td>   12.997</td> <td> 0.000</td> <td>    0.002</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>('Adj Close', '005930.KS')</th> <td>    0.0190</td> <td>    0.002</td> <td>    8.846</td> <td> 0.000</td> <td>    0.015</td> <td>    0.023</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>('Volume', '035420.KS')</th>    <td> 4.902e-05</td> <td> 1.42e-05</td> <td>    3.464</td> <td> 0.002</td> <td> 2.02e-05</td> <td> 7.79e-05</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td> 1.774</td> <th>  Durbin-Watson:     </th> <td>   0.875</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th> <td> 0.412</td> <th>  Jarque-Bera (JB):  </th> <td>   1.618</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>          <td> 0.422</td> <th>  Prob(JB):          </th> <td>   0.445</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>      <td> 2.369</td> <th>  Cond. No.          </th> <td>4.67e+07</td>\n",
       "</tr>\n",
       "</table><br/><br/>Notes:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 4.67e+07. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:                  ^KS11   R-squared:                       0.887\n",
       "Model:                            OLS   Adj. R-squared:                  0.876\n",
       "Method:                 Least Squares   F-statistic:                     81.12\n",
       "Date:                Wed, 30 Mar 2022   Prob (F-statistic):           9.03e-15\n",
       "Time:                        16:51:29   Log-Likelihood:                -146.38\n",
       "No. Observations:                  35   AIC:                             300.8\n",
       "Df Residuals:                      31   BIC:                             307.0\n",
       "Df Model:                           3                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "==============================================================================================\n",
       "                                 coef    std err          t      P>|t|      [0.025      0.975]\n",
       "----------------------------------------------------------------------------------------------\n",
       "const                        734.6837    187.354      3.921      0.000     352.572    1116.795\n",
       "('Low', '035420.KS')           0.0020      0.000     12.997      0.000       0.002       0.002\n",
       "('Adj Close', '005930.KS')     0.0190      0.002      8.846      0.000       0.015       0.023\n",
       "('Volume', '035420.KS')     4.902e-05   1.42e-05      3.464      0.002    2.02e-05    7.79e-05\n",
       "==============================================================================\n",
       "Omnibus:                        1.774   Durbin-Watson:                   0.875\n",
       "Prob(Omnibus):                  0.412   Jarque-Bera (JB):                1.618\n",
       "Skew:                           0.422   Prob(JB):                        0.445\n",
       "Kurtosis:                       2.369   Cond. No.                     4.67e+07\n",
       "==============================================================================\n",
       "\n",
       "Notes:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "[2] The condition number is large, 4.67e+07. This might indicate that there are\n",
       "strong multicollinearity or other numerical problems.\n",
       "\"\"\""
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 최종 모형 도출\n",
    "X = sm.add_constant(stock_price[forward_valriables])\n",
    "model = sm.OLS(y, X).fit(disp=0)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 스텝별로 조정된 결정계수를 출력하는 표 시각화\n",
    "fig = plt.figure(figsize=(10,10))\n",
    "plt.plot(steps, adj_r_squared_list, marker='o')\n",
    "plt.ylabel('adj_r_squared',fontsize=15)\n",
    "plt.grid(True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                  ^KS11   R-squared:                       0.887\n",
      "Model:                            OLS   Adj. R-squared:                  0.876\n",
      "Method:                 Least Squares   F-statistic:                     81.12\n",
      "Date:                Wed, 30 Mar 2022   Prob (F-statistic):           9.03e-15\n",
      "Time:                        16:51:34   Log-Likelihood:                -146.38\n",
      "No. Observations:                  35   AIC:                             300.8\n",
      "Df Residuals:                      31   BIC:                             307.0\n",
      "Df Model:                           3                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================================\n",
      "                                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "----------------------------------------------------------------------------------------------\n",
      "const                        734.6837    187.354      3.921      0.000     352.572    1116.795\n",
      "('Low', '035420.KS')           0.0020      0.000     12.997      0.000       0.002       0.002\n",
      "('Adj Close', '005930.KS')     0.0190      0.002      8.846      0.000       0.015       0.023\n",
      "('Volume', '035420.KS')     4.902e-05   1.42e-05      3.464      0.002    2.02e-05    7.79e-05\n",
      "==============================================================================\n",
      "Omnibus:                        1.774   Durbin-Watson:                   0.875\n",
      "Prob(Omnibus):                  0.412   Jarque-Bera (JB):                1.618\n",
      "Skew:                           0.422   Prob(JB):                        0.445\n",
      "Kurtosis:                       2.369   Cond. No.                     4.67e+07\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 4.67e+07. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def forward_feature_selection(feature, target, sl_enter=.05, sl_remove=.05):\n",
    "    ## 전진 단계별 선택법\n",
    "\n",
    "    import statsmodels.api as sm\n",
    "    import pandas as pd\n",
    "    import matplotlib.pyplot as plt\n",
    "    import warnings\n",
    "    warnings.filterwarnings(\"ignore\")\n",
    "    \n",
    "    # 변수 설정\n",
    "    variables = feature.columns.tolist() \n",
    "    # 선택된 변수들 list 생성\n",
    "    forward_valriables = []\n",
    "    # 각 스텝별로 선택된 변수들\n",
    "    sv_per_step = [] \n",
    "    # 각 스텝별 수정된 결정계수\n",
    "    adj_r_squared_list = []\n",
    "    # 스텝\n",
    "    steps = []\n",
    "    step = 0\n",
    "\n",
    "    while len(variables) > 0:\n",
    "        remainder = list(set(variables) - set(forward_valriables))\n",
    "        pval = pd.Series(index=remainder) ## 변수의 p-value\n",
    "        ## 기존에 포함된 변수와 새로운 변수 하나씩 돌아가면서 \n",
    "        ## 선형 모형을 적합한다.\n",
    "        for col in remainder: \n",
    "            X = feature[forward_valriables+[col]]\n",
    "            X = sm.add_constant(X)\n",
    "            model = sm.OLS(target,X).fit(disp=0)\n",
    "            pval[col] = model.pvalues[col]\n",
    "    \n",
    "        min_pval = pval.min()\n",
    "        if min_pval < sl_enter: ## 최소 p-value 값이 기준 값보다 작으면 포함\n",
    "            forward_valriables.append(pval.idxmin())\n",
    "            ## 선택된 변수들에대해서\n",
    "            ## 어떤 변수를 제거할지 고른다.\n",
    "            while len(forward_valriables) > 0:\n",
    "                selected_X = feature[forward_valriables]\n",
    "                selected_X = sm.add_constant(selected_X)\n",
    "                selected_pval = sm.OLS(target,selected_X).fit(disp=0).pvalues[1:] ## 절편항의 p-value는 뺀다\n",
    "                max_pval = selected_pval.max()\n",
    "                if max_pval >= sl_remove: ## 최대 p-value값이 기준값보다 크거나 같으면 제외\n",
    "                    remove_variable = selected_pval.idxmax()\n",
    "                    forward_valriables.remove(remove_variable)\n",
    "                else:\n",
    "                    break\n",
    "            \n",
    "            step += 1\n",
    "            steps.append(step)\n",
    "            adj_r_squared = sm.OLS(target,sm.add_constant(feature[forward_valriables])).fit(disp=0).rsquared_adj\n",
    "            adj_r_squared_list.append(adj_r_squared)\n",
    "            sv_per_step.append(forward_valriables)\n",
    "        else:\n",
    "            break\n",
    "\n",
    "    # 최종 모형 도출\n",
    "    model = sm.OLS(target, sm.add_constant(stock_price[forward_valriables])).fit(disp=0)\n",
    "    print(model.summary())\n",
    "\n",
    "    # 스텝별로 조정된 결정계수를 출력하는 표 시각화\n",
    "    fig = plt.figure(figsize=(10,10))\n",
    "    plt.plot(steps, adj_r_squared_list, marker='o')\n",
    "    plt.ylabel('adj_r_squared',fontsize=15)\n",
    "    plt.grid(True)\n",
    "    plt.show()\n",
    "    \n",
    "    return model\n",
    "    \n",
    "forward_model = forward_feature_selection(feature, target)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3-2 후진제거법"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                  ^KS11   R-squared:                       0.977\n",
      "Model:                            OLS   Adj. R-squared:                  0.962\n",
      "Method:                 Least Squares   F-statistic:                     67.22\n",
      "Date:                Wed, 30 Mar 2022   Prob (F-statistic):           4.21e-14\n",
      "Time:                        16:51:37   Log-Likelihood:                -118.88\n",
      "No. Observations:                  35   AIC:                             265.8\n",
      "Df Residuals:                      21   BIC:                             287.5\n",
      "Df Model:                          13                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================================\n",
      "                                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "----------------------------------------------------------------------------------------------\n",
      "const                        716.5264    181.792      3.941      0.001     338.470    1094.583\n",
      "('Adj Close', '005930.KS')     1.4340      0.457      3.136      0.005       0.483       2.385\n",
      "('Adj Close', '035420.KS')    -1.0009      0.329     -3.042      0.006      -1.685      -0.317\n",
      "('Close', '005930.KS')        -1.4244      0.456     -3.122      0.005      -2.373      -0.476\n",
      "('Close', '035420.KS')         1.0028      0.329      3.050      0.006       0.319       1.687\n",
      "('Close', '035720.KS')        -0.0066      0.003     -2.444      0.023      -0.012      -0.001\n",
      "('High', '000660.KS')         -0.0156      0.003     -6.150      0.000      -0.021      -0.010\n",
      "('High', '005930.KS')          0.0154      0.007      2.293      0.032       0.001       0.029\n",
      "('High', '035420.KS')         -0.0032      0.001     -3.678      0.001      -0.005      -0.001\n",
      "('Low', '000660.KS')           0.0141      0.002      6.713      0.000       0.010       0.018\n",
      "('Low', '035420.KS')           0.0043      0.001      3.794      0.001       0.002       0.007\n",
      "('Open', '035420.KS')         -0.0019      0.001     -2.129      0.045      -0.004   -4.42e-05\n",
      "('Open', '035720.KS')          0.0074      0.002      3.781      0.001       0.003       0.012\n",
      "('Volume', '035420.KS')     8.509e-05   1.38e-05      6.151      0.000    5.63e-05       0.000\n",
      "==============================================================================\n",
      "Omnibus:                        0.642   Durbin-Watson:                   2.019\n",
      "Prob(Omnibus):                  0.725   Jarque-Bera (JB):                0.677\n",
      "Skew:                           0.289   Prob(JB):                        0.713\n",
      "Kurtosis:                       2.638   Cond. No.                     1.21e+08\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 1.21e+08. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def backward_regression(feature, target,\n",
    "                           initial_list=[], \n",
    "                           threshold_out = 0.05):\n",
    "    \n",
    "    import matplotlib.pyplot as plt\n",
    "    import statsmodels.api as sm\n",
    "    import pandas as pd\n",
    "    import warnings\n",
    "    warnings.filterwarnings(\"ignore\")\n",
    "    \n",
    "    feature_list = feature.columns.tolist()  \n",
    "    sv_per_step = [] ## 각 스텝별로 선택된 변수들\n",
    "    adj_r_squared_list = [] ## 각 스텝별 수정된 결정계수\n",
    "    steps = [] ## 스텝\n",
    "    step = 0\n",
    "    included = feature.columns.tolist() \n",
    "    \n",
    "    while True:\n",
    "        changed=False\n",
    "        model = sm.OLS(target, sm.add_constant(feature[included])).fit(disp=0)\n",
    "        # use all coefs except intercept\n",
    "        pvalues = model.pvalues.iloc[1:] # 각 feature의 P값을 의미함\n",
    "        worst_pval = pvalues.max()\t# P 값이 가장 높은 것 선정\n",
    "        if worst_pval > threshold_out:\n",
    "            changed=True\n",
    "            worst_feature = pvalues.idxmax()\n",
    "            included.remove(worst_feature)\n",
    "        \n",
    "        step += 1\n",
    "        steps.append(step)        \n",
    "        adj_r_squared = sm.OLS(target, sm.add_constant(feature[included])).fit(disp=0).rsquared_adj\n",
    "        adj_r_squared_list.append(adj_r_squared)\n",
    "        sv_per_step.append(included.copy())\n",
    "        if not changed:\n",
    "            break\n",
    "    \n",
    "    # 최종모델 출력\n",
    "    model = sm.OLS(target, sm.add_constant(feature[included])).fit()\n",
    "    print(model.summary())\n",
    "\n",
    "    \n",
    "    fig = plt.figure(figsize=(10,10))\n",
    "    plt.plot(steps, adj_r_squared_list, marker='o')\n",
    "    plt.ylabel('adj_r_squared',fontsize=15)\n",
    "    plt.grid(True)\n",
    "    plt.show()\n",
    "\n",
    "    return model\n",
    "\n",
    "backward_model = backward_regression(feature, target)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3-3 단계적 선택법"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                  ^KS11   R-squared:                       0.887\n",
      "Model:                            OLS   Adj. R-squared:                  0.876\n",
      "Method:                 Least Squares   F-statistic:                     81.12\n",
      "Date:                Wed, 30 Mar 2022   Prob (F-statistic):           9.03e-15\n",
      "Time:                        16:51:40   Log-Likelihood:                -146.38\n",
      "No. Observations:                  35   AIC:                             300.8\n",
      "Df Residuals:                      31   BIC:                             307.0\n",
      "Df Model:                           3                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================================\n",
      "                                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "----------------------------------------------------------------------------------------------\n",
      "const                        734.6837    187.354      3.921      0.000     352.572    1116.795\n",
      "('Low', '035420.KS')           0.0020      0.000     12.997      0.000       0.002       0.002\n",
      "('Adj Close', '005930.KS')     0.0190      0.002      8.846      0.000       0.015       0.023\n",
      "('Volume', '035420.KS')     4.902e-05   1.42e-05      3.464      0.002    2.02e-05    7.79e-05\n",
      "==============================================================================\n",
      "Omnibus:                        1.774   Durbin-Watson:                   0.875\n",
      "Prob(Omnibus):                  0.412   Jarque-Bera (JB):                1.618\n",
      "Skew:                           0.422   Prob(JB):                        0.445\n",
      "Kurtosis:                       2.369   Cond. No.                     4.67e+07\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The condition number is large, 4.67e+07. This might indicate that there are\n",
      "strong multicollinearity or other numerical problems.\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def stepwise_feature_selection(feature, target, variables=feature.columns.tolist()):\n",
    "    import statsmodels.api as sm\n",
    "    import matplotlib.pyplot as plt\n",
    "    import warnings\n",
    "    import pandas as pd\n",
    "    warnings.filterwarnings(\"ignore\")\n",
    "    \n",
    "    y = target ## 반응 변수\n",
    "\n",
    "    selected_variables = [] ## 선택된 변수들\n",
    "    sl_enter = 0.05\n",
    "    sl_remove = 0.05\n",
    "    \n",
    "    sv_per_step = [] ## 각 스텝별로 선택된 변수들\n",
    "    adjusted_r_squared = [] ## 각 스텝별 수정된 결정계수\n",
    "    steps = [] ## 스텝\n",
    "    step = 0\n",
    "    while len(variables) > 0:\n",
    "        remainder = list(set(variables) - set(selected_variables))\n",
    "        pval = pd.Series(index=remainder) ## 변수의 p-value\n",
    "        ## 기존에 포함된 변수와 새로운 변수 하나씩 돌아가면서 \n",
    "        ## 선형 모형을 적합한다.\n",
    "        for col in remainder: \n",
    "            X = feature[selected_variables+[col]]\n",
    "            X = sm.add_constant(X)\n",
    "            model = sm.OLS(y,X).fit(disp=0)\n",
    "            pval[col] = model.pvalues[col]\n",
    "    \n",
    "        min_pval = pval.min()\n",
    "        if min_pval < sl_enter: ## 최소 p-value 값이 기준 값보다 작으면 포함\n",
    "            selected_variables.append(pval.idxmin())\n",
    "            ## 선택된 변수들에대해서\n",
    "            ## 어떤 변수를 제거할지 고른다.\n",
    "            while len(selected_variables) > 0:\n",
    "                selected_X = feature[selected_variables]\n",
    "                selected_X = sm.add_constant(selected_X)\n",
    "                selected_pval = sm.OLS(y,selected_X).fit(disp=0).pvalues[1:] ## 절편항의 p-value는 뺀다\n",
    "                max_pval = selected_pval.max()\n",
    "                if max_pval >= sl_remove: ## 최대 p-value값이 기준값보다 크거나 같으면 제외\n",
    "                    remove_variable = selected_pval.idxmax()\n",
    "                    selected_variables.remove(remove_variable)\n",
    "                else:\n",
    "                    break\n",
    "            \n",
    "            step += 1\n",
    "            steps.append(step)\n",
    "            adj_r_squared = sm.OLS(y,sm.add_constant(feature[selected_variables])).fit(disp=0).rsquared_adj\n",
    "            adjusted_r_squared.append(adj_r_squared)\n",
    "            sv_per_step.append(selected_variables.copy())\n",
    "        else:\n",
    "            break\n",
    "    \n",
    "    model = sm.OLS(y, sm.add_constant(feature[selected_variables])).fit()\n",
    "    print(model.summary())\n",
    "    \n",
    "    # 시각화\n",
    "    fig = plt.figure(figsize=(10,10))\n",
    "    plt.plot(steps,adjusted_r_squared, marker='o')\n",
    "    plt.ylabel('Adjusted R Squared',fontsize=15)\n",
    "    plt.grid(True)\n",
    "    plt.show()\n",
    "\n",
    "    return model\n",
    "\n",
    "stepwise_model = stepwise_feature_selection(feature, target)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
  },
  "kernelspec": {
   "display_name": "Python 3.9.4 64-bit",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}