conjointAnalysis_main.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a5c33df4",
   "metadata": {},
   "source": [
    "# コンジョイント分析を用いて、消費者の好みを分析してみよう  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6602ef7c",
   "metadata": {},
   "source": [
    "以下は、コンジョイント分析を行うためのpythonコードです。\n",
    "- pythonのバージョンは3.8  \n",
    "- 他のモジュールのバージョンについては、githubにアップロードしているファイルをダウンロードして、`myenv.txt` をご確認ください。\n",
    "\n",
    "## モジュールのインポート"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7d91741c",
   "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 statsmodels.api as sm"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2411f45b",
   "metadata": {},
   "source": [
    "## データの読み込み"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f3255d7d",
   "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>auto-lock</th>\n",
       "      <th>distToUniv</th>\n",
       "      <th>isParking</th>\n",
       "      <th>fee</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    auto-lock  distToUniv  isParking  fee\n",
       "0           1           1          1    2\n",
       "1           1           1          2    1\n",
       "2           1           2          1    1\n",
       "3           1           2          2    3\n",
       "4           2           2          2    2\n",
       "5           2           2          1    1\n",
       "6           2           1          2    1\n",
       "7           2           1          1    3\n",
       "8           1           1          1    2\n",
       "9           1           1          2    1\n",
       "10          1           2          1    1\n",
       "11          1           2          2    3\n",
       "12          2           2          2    2\n",
       "13          2           2          1    1\n",
       "14          2           1          2    1\n",
       "15          2           1          1    3"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('./data/data.csv')\n",
    "# x, y の指定\n",
    "y = pd.DataFrame(df['score'])\n",
    "x = df.drop(columns=['score'])\n",
    "x.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a023f2e",
   "metadata": {},
   "source": [
    "このデータは２人分、合計１６個の回答を示しています。例えば一番上の`auto-lock 1, distToUniv 1, isParking 1, fee 2`では、  \n",
    "オートロックあり、大学からの距離がちかい、駐車場あり、家賃８万（１～３があって、それぞれ、６，８，１０万に対応）という条件を示しています。これに対して、それぞれ被験者（今回は私が勝手に回答）が魅力度のスコアを付けます。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9838c23d",
   "metadata": {},
   "source": [
    "## ダミー変数への変換  \n",
    "ここでは、`pd.get_dummies`関数を用いて、そのカードに、該当の項目が書かれているかどうかを 0/1で示します。 \n",
    "one-hotベクトルに直しているのと似ています。  \n",
    "`drop_first`をtrueにして、その項目のはじめの要素は削除するようにしています。例えば、駐車場の有無では、駐車場がある、という要素が0である場合と、駐車場がない、という要素が1である場合は同じ意味です。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d8faf57a",
   "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>auto-lock_2</th>\n",
       "      <th>distToUniv_2</th>\n",
       "      <th>isParking_2</th>\n",
       "      <th>fee_2</th>\n",
       "      <th>fee_3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   auto-lock_2  distToUniv_2  isParking_2  fee_2  fee_3\n",
       "0            0             0            0      1      0\n",
       "1            0             0            1      0      0\n",
       "2            0             1            0      0      0\n",
       "3            0             1            1      0      1\n",
       "4            1             1            1      1      0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_dum = pd.get_dummies(x, columns=x.columns, drop_first=True)\n",
    "x_dum.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "725ee6e9",
   "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>auto-lock_1</th>\n",
       "      <th>auto-lock_2</th>\n",
       "      <th>distToUniv_1</th>\n",
       "      <th>distToUniv_2</th>\n",
       "      <th>isParking_1</th>\n",
       "      <th>isParking_2</th>\n",
       "      <th>fee_1</th>\n",
       "      <th>fee_2</th>\n",
       "      <th>fee_3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   auto-lock_1  auto-lock_2  distToUniv_1  distToUniv_2  isParking_1  \\\n",
       "0            1            0             1             0            1   \n",
       "1            1            0             1             0            0   \n",
       "2            1            0             0             1            1   \n",
       "3            1            0             0             1            0   \n",
       "4            0            1             0             1            0   \n",
       "\n",
       "   isParking_2  fee_1  fee_2  fee_3  \n",
       "0            0      0      1      0  \n",
       "1            1      1      0      0  \n",
       "2            0      1      0      0  \n",
       "3            1      0      0      1  \n",
       "4            1      0      1      0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# drop_firstを無効にした場合を確認。_1のものが残っていることがわかる。ここではこのデータは使わない\n",
    "x_dum_noDrop = pd.get_dummies(x, columns=x.columns, drop_first=False)\n",
    "x_dum_noDrop.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9db87dfe",
   "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>score</th>\n",
       "      <th>auto-lock</th>\n",
       "      <th>distToUniv</th>\n",
       "      <th>isParking</th>\n",
       "      <th>fee</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>7.187500</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.750000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2.644964</td>\n",
       "      <td>0.516398</td>\n",
       "      <td>0.516398</td>\n",
       "      <td>0.516398</td>\n",
       "      <td>0.856349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>6.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>7.000000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>8.625000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>12.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           score  auto-lock  distToUniv  isParking        fee\n",
       "count  16.000000  16.000000   16.000000  16.000000  16.000000\n",
       "mean    7.187500   1.500000    1.500000   1.500000   1.750000\n",
       "std     2.644964   0.516398    0.516398   0.516398   0.856349\n",
       "min     2.000000   1.000000    1.000000   1.000000   1.000000\n",
       "25%     6.000000   1.000000    1.000000   1.000000   1.000000\n",
       "50%     7.000000   1.500000    1.500000   1.500000   1.500000\n",
       "75%     8.625000   2.000000    2.000000   2.000000   2.250000\n",
       "max    12.000000   2.000000    2.000000   2.000000   3.000000"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe() #要素の平均や標準偏差などの基本的な統計データを表示させる"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53dda34a",
   "metadata": {},
   "source": [
    "## 切片を追加 \n",
    "また、コンジョイントカードに記載されている内容に加えて、その他の影響がある場合に備えて、定数項も加えます。\n",
    "\n",
    "この操作によってフィッティングするときの切片を計算することができます。  \n",
    "https://www.statsmodels.org/stable/generated/statsmodels.tools.tools.add_constant.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "8ca5913f",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\itaku\\anaconda3\\envs\\py38_geopanda\\lib\\site-packages\\statsmodels\\tsa\\tsatools.py:142: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
      "  x = pd.concat(x[::order], 1)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>const</th>\n",
       "      <th>auto-lock_2</th>\n",
       "      <th>distToUniv_2</th>\n",
       "      <th>isParking_2</th>\n",
       "      <th>fee_2</th>\n",
       "      <th>fee_3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   const  auto-lock_2  distToUniv_2  isParking_2  fee_2  fee_3\n",
       "0    1.0            0             0            0      1      0\n",
       "1    1.0            0             0            1      0      0\n",
       "2    1.0            0             1            0      0      0\n",
       "3    1.0            0             1            1      0      1\n",
       "4    1.0            1             1            1      1      0\n",
       "5    1.0            1             1            0      0      0\n",
       "6    1.0            1             0            1      0      0\n",
       "7    1.0            1             0            0      0      1\n",
       "8    1.0            0             0            0      1      0\n",
       "9    1.0            0             0            1      0      0"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_dum=sm.add_constant(x_dum) # constという要素を追加\n",
    "x_dum.head(10) # constが追加されたことを確認"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8470427d",
   "metadata": {},
   "source": [
    "## OLS (Ordinary Least Squares) でフィッティング"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cbabd598",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\itaku\\anaconda3\\envs\\py38_geopanda\\lib\\site-packages\\scipy\\stats\\stats.py:1541: UserWarning: kurtosistest only valid for n>=20 ... continuing anyway, n=16\n",
      "  warnings.warn(\"kurtosistest only valid for n>=20 ... continuing \"\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>          <td>score</td>      <th>  R-squared:         </th> <td>   0.959</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.938</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   46.32</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Mon, 03 Jan 2022</td> <th>  Prob (F-statistic):</th> <td>1.35e-06</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>13:46:07</td>     <th>  Log-Likelihood:    </th> <td> -12.272</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td>    16</td>      <th>  AIC:               </th> <td>   36.54</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td>    10</td>      <th>  BIC:               </th> <td>   41.18</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>     5</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>    9.7500</td> <td>    0.368</td> <td>   26.463</td> <td> 0.000</td> <td>    8.929</td> <td>   10.571</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>auto-lock_2</th>  <td>    0.7500</td> <td>    0.330</td> <td>    2.276</td> <td> 0.046</td> <td>    0.016</td> <td>    1.484</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>distToUniv_2</th> <td>   -3.0000</td> <td>    0.330</td> <td>   -9.104</td> <td> 0.000</td> <td>   -3.734</td> <td>   -2.266</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isParking_2</th>  <td>    0.3750</td> <td>    0.330</td> <td>    1.138</td> <td> 0.282</td> <td>   -0.359</td> <td>    1.109</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>fee_2</th>        <td>   -1.6875</td> <td>    0.404</td> <td>   -4.181</td> <td> 0.002</td> <td>   -2.587</td> <td>   -0.788</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>fee_3</th>        <td>   -4.8125</td> <td>    0.404</td> <td>  -11.924</td> <td> 0.000</td> <td>   -5.712</td> <td>   -3.913</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td> 0.797</td> <th>  Durbin-Watson:     </th> <td>   2.795</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th> <td> 0.671</td> <th>  Jarque-Bera (JB):  </th> <td>   0.028</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>          <td>-0.010</td> <th>  Prob(JB):          </th> <td>   0.986</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>      <td> 3.205</td> <th>  Cond. No.          </th> <td>    4.54</td>\n",
       "</tr>\n",
       "</table><br/><br/>Notes:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:                  score   R-squared:                       0.959\n",
       "Model:                            OLS   Adj. R-squared:                  0.938\n",
       "Method:                 Least Squares   F-statistic:                     46.32\n",
       "Date:                Mon, 03 Jan 2022   Prob (F-statistic):           1.35e-06\n",
       "Time:                        13:46:07   Log-Likelihood:                -12.272\n",
       "No. Observations:                  16   AIC:                             36.54\n",
       "Df Residuals:                      10   BIC:                             41.18\n",
       "Df Model:                           5                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "================================================================================\n",
       "                   coef    std err          t      P>|t|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------\n",
       "const            9.7500      0.368     26.463      0.000       8.929      10.571\n",
       "auto-lock_2      0.7500      0.330      2.276      0.046       0.016       1.484\n",
       "distToUniv_2    -3.0000      0.330     -9.104      0.000      -3.734      -2.266\n",
       "isParking_2      0.3750      0.330      1.138      0.282      -0.359       1.109\n",
       "fee_2           -1.6875      0.404     -4.181      0.002      -2.587      -0.788\n",
       "fee_3           -4.8125      0.404    -11.924      0.000      -5.712      -3.913\n",
       "==============================================================================\n",
       "Omnibus:                        0.797   Durbin-Watson:                   2.795\n",
       "Prob(Omnibus):                  0.671   Jarque-Bera (JB):                0.028\n",
       "Skew:                          -0.010   Prob(JB):                        0.986\n",
       "Kurtosis:                       3.205   Cond. No.                         4.54\n",
       "==============================================================================\n",
       "\n",
       "Notes:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "\"\"\""
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = sm.OLS(y, x_dum)\n",
    "# フィッティングを実行\n",
    "result = model.fit()\n",
    "# 結果の一覧を表示\n",
    "result.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa44c912",
   "metadata": {},
   "source": [
    "## 結果の一部を取り出し  \n",
    "`結果を見てみる`セクションで議論するため、weightとｐ値を取り出します"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f158f820",
   "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>weight</th>\n",
       "      <th>p_val</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>const</th>\n",
       "      <td>9.7500</td>\n",
       "      <td>1.369002e-10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>auto-lock_2</th>\n",
       "      <td>0.7500</td>\n",
       "      <td>4.610362e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>distToUniv_2</th>\n",
       "      <td>-3.0000</td>\n",
       "      <td>3.732512e-06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isParking_2</th>\n",
       "      <td>0.3750</td>\n",
       "      <td>2.816648e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_2</th>\n",
       "      <td>-1.6875</td>\n",
       "      <td>1.884264e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_3</th>\n",
       "      <td>-4.8125</td>\n",
       "      <td>3.101164e-07</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              weight         p_val\n",
       "const         9.7500  1.369002e-10\n",
       "auto-lock_2   0.7500  4.610362e-02\n",
       "distToUniv_2 -3.0000  3.732512e-06\n",
       "isParking_2   0.3750  2.816648e-01\n",
       "fee_2        -1.6875  1.884264e-03\n",
       "fee_3        -4.8125  3.101164e-07"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_result_selected = pd.DataFrame({\n",
    "    'weight': result.params.values\n",
    "    , 'p_val': result.pvalues\n",
    "})\n",
    "df_result_selected.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba4ac286",
   "metadata": {},
   "source": [
    "## 可視化のための準備  \n",
    "コンジョイント分析の結果を可視化するための準備を行います。上のdrop_Firstを有効にして、_ 1 がつく変数は落とされていました。グラフで表示させるため復帰させます。これらの重みは0とします。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b0e0c2c7",
   "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>weight</th>\n",
       "      <th>p_val</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>const</th>\n",
       "      <td>9.7500</td>\n",
       "      <td>1.369002e-10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>auto-lock_2</th>\n",
       "      <td>0.7500</td>\n",
       "      <td>4.610362e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>distToUniv_2</th>\n",
       "      <td>-3.0000</td>\n",
       "      <td>3.732512e-06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isParking_2</th>\n",
       "      <td>0.3750</td>\n",
       "      <td>2.816648e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_2</th>\n",
       "      <td>-1.6875</td>\n",
       "      <td>1.884264e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_3</th>\n",
       "      <td>-4.8125</td>\n",
       "      <td>3.101164e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>auto-lock_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>distToUniv_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isParking_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              weight         p_val\n",
       "const         9.7500  1.369002e-10\n",
       "auto-lock_2   0.7500  4.610362e-02\n",
       "distToUniv_2 -3.0000  3.732512e-06\n",
       "isParking_2   0.3750  2.816648e-01\n",
       "fee_2        -1.6875  1.884264e-03\n",
       "fee_3        -4.8125  3.101164e-07\n",
       "auto-lock_1   0.0000  0.000000e+00\n",
       "distToUniv_1  0.0000  0.000000e+00\n",
       "isParking_1   0.0000  0.000000e+00\n",
       "fee_1         0.0000  0.000000e+00"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for s in df_result_selected.index:\n",
    "    partitioned_string = s.partition('_')\n",
    "    if partitioned_string[2] == \"2\":\n",
    "        valBase = partitioned_string[0] + \"_1\"\n",
    "        df_valBase = pd.DataFrame(data =np.zeros((1,2)),\n",
    "                                  index = [valBase],\n",
    "                                  columns = [\"weight\",\"p_val\"])\n",
    "        df_result_selected = pd.concat([df_result_selected,df_valBase])\n",
    "\n",
    "df_result_selected.head(20) "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "82886f8c",
   "metadata": {},
   "source": [
    "## ｐ値に応じてバーの色を変える：  \n",
    "p値が0.01以上、0.05以下の場合はシアン、0.01以下の場合は青、0.05以上（あまり信用できない）の場合は赤に設定します。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "0f7be610",
   "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>weight</th>\n",
       "      <th>p_val</th>\n",
       "      <th>bar_col</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>const</th>\n",
       "      <td>9.7500</td>\n",
       "      <td>1.369002e-10</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>auto-lock_2</th>\n",
       "      <td>0.7500</td>\n",
       "      <td>4.610362e-02</td>\n",
       "      <td>Cyan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>distToUniv_2</th>\n",
       "      <td>-3.0000</td>\n",
       "      <td>3.732512e-06</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isParking_2</th>\n",
       "      <td>0.3750</td>\n",
       "      <td>2.816648e-01</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_2</th>\n",
       "      <td>-1.6875</td>\n",
       "      <td>1.884264e-03</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_3</th>\n",
       "      <td>-4.8125</td>\n",
       "      <td>3.101164e-07</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>auto-lock_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>distToUniv_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isParking_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fee_1</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              weight         p_val bar_col\n",
       "const         9.7500  1.369002e-10    blue\n",
       "auto-lock_2   0.7500  4.610362e-02    Cyan\n",
       "distToUniv_2 -3.0000  3.732512e-06    blue\n",
       "isParking_2   0.3750  2.816648e-01     red\n",
       "fee_2        -1.6875  1.884264e-03    blue\n",
       "fee_3        -4.8125  3.101164e-07    blue\n",
       "auto-lock_1   0.0000  0.000000e+00    blue\n",
       "distToUniv_1  0.0000  0.000000e+00    blue\n",
       "isParking_1   0.0000  0.000000e+00    blue\n",
       "fee_1         0.0000  0.000000e+00    blue"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bar_col = []\n",
    "for p_val in df_result_selected['p_val']:\n",
    "    # print(p_val)\n",
    "    if 0.01 < p_val < 0.05:\n",
    "        bar_col.append('Cyan')\n",
    "    elif p_val < 0.01:\n",
    "        bar_col.append('blue')\n",
    "    else:\n",
    "        bar_col.append('red')\n",
    "\n",
    "# p値が0.05以下のものを青、そうでないものを青とする\n",
    "df_bar_col = pd.DataFrame(data = bar_col, \n",
    "                          columns=['bar_col'], \n",
    "                          index = df_result_selected.index)\n",
    "\n",
    "df_result_selected = pd.concat([df_result_selected,df_bar_col], axis=1)\n",
    "df_result_selected.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78eff107",
   "metadata": {},
   "source": [
    "## グラフを表示させる  \n",
    "_ 1 とつくものが基準になっているので、それをもとに同一カテゴリが正または負の影響があるか見てください。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "614293d8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# プロットするときに日本語でも文字化けしないように設定\n",
    "from matplotlib import rcParams\n",
    "plt.rcParams[\"font.family\"] = \"MS Gothic\"\n",
    "# アルファベット順位\n",
    "df_result_selected = df_result_selected.sort_index()\n",
    "xbar = np.arange(len(df_result_selected['weight']))\n",
    "plt.barh(xbar, df_result_selected['weight'], color=df_result_selected['bar_col'])\n",
    "index_JP = [\"駐車場なし\",\"駐車場あり\",\"家賃１０万\",\"家賃８万\",\"家賃６万\",\"大学から遠い\",\"大学から近い\",\"定数項\",\"オートロックなし\",\"オートロックあり\"]\n",
    "plt.yticks(xbar, labels=index_JP[::-1]) # 順番があうように順番を逆にする\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0e30d5d2",
   "metadata": {},
   "source": [
    "# 結果を見てみる  \n",
    "\n",
    "## 上のOLS Regression Resultsについて（result.summary()で表示したところ）\n",
    "- R-squaredは約0.96と高い値を示している\n",
    "=> 良いフィッティング結果を得ることができた。簡単なデータではあったが、スコアを決定する要因の全体の9割以上を説明することができている。私の重視する家賃、大学までの近さがともに入っているためだと考えられる。ただ、これが私の家選びの傾向を完全に理解した、ということにはならないと思います（「考えたこと」章を参照ください）\n",
    "- P>|t|\tは、isParking_2以外、統計的に有意である（有意水準を5%とした場合）ことがわかります。\n",
    "=> isParkingは個人的にどちらでも良いので、私がこの練習用データを作るときは、ほとんど見ずに回答していました。そのため、この限られたデータではうまくフィッティングできなかったのではないかと考えられます。もしより多くの似た考えをもつ回答者がいれば、このｐ値もより小さく（有意になり）、かつ重みが小さな値に収束していくはずです。P>|t|の欄の値が大きいものばかりだと、たまたま重みが大きくなっただけの偶然であることが否定できず、コンジョイント分析から多くの示唆を得ることができないため、今回はよい練習データになっていてよかったです。\n",
    "- 多重共線性について：Cond. No.(Condition number)を確認します。statsmodelsのドキュメンテーションによると、  \n",
    "\n",
    "> One way to assess multicollinearity is to compute the condition number. Values over 20 are worrisome (see Greene 4.9).  \n",
    "\n",
    "とあります。今回は20以下なので、多重共線性についてもひとまず大丈夫そうです。\n",
    "\n",
    "https://www.statsmodels.org/stable/examples/notebooks/generated/ols.html\n",
    "\n",
    "- 最後の色に分けられたグラフでは、**私が家賃安め＆大学から近め を優先してよいスコアを付けたので、納得のいくグラフ**です。駐車場がない場合が少しポジティブな方向に出ていますが、特に私は考えず記入していました。ここでは、ｐ値が0.05より高く赤で示されていて、特に考えず記入しなかったことともつじつまがあっています。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "115733e1",
   "metadata": {},
   "source": [
    "# 考えたこと  \n",
    "この分析を勉強してみて考えたことを以下に記載します。この分析は勉強中なので的外れなところもあるかもしれませんのでご注意ください。\n",
    "## コンジョイント分析全体について  \n",
    "- 今回はpythonによる実装を中心に議論しましたが、解析の中身自体は重回帰をして、その重みをもって議論しているため、比較的シンプルな方法なのではないかと思いました。  \n",
    "- 機械学習でいう、LIMEとアイデア似ているところが多いな思いました。LIMEでは、例えば画像の場合、ランダムにブラックアウトさせて、そのときのスコアの変動を見ます。そこでも線形回帰が使われていて、その重みを用いて機械学習や深層学習による判断根拠の可視化を行います。  \n",
    "\n",
    "## 疑問点や改良点について  \n",
    "- 今回は、被験者の少ないデータを自作し、テストしています。しかし、被験者が多くなった場合は、スコアのベースラインも被験者によって異なるのでその補正が必要だと思います。例えば、被験者によって、高めに点を付ける人、そうでない人が存在すると、各被験者のスコアの平均で割り算したり、何らかの標準化が必要です。  \n",
    "- 重回帰による重要度の議論について：上で述べたLIMEでは、線形回帰をするために重回帰を用いたり、決定木を用いています。コンジョイント分析でも、うまく重回帰ではフィッティングできず、決定係数の低い場合は決定木を用いてみても良いのかもしれませんね。\n",
    "- 今回の練習データでは決定係数も高く、P>|t|の値も良好でした。ただ、この結果から私の家選びの基準を完全に推論できるかというとそうでもなくて、例えば、家がキレイか、ユニットバス/セパレートか、なども個人的に重視するポイントです。確かに家賃も気にはするものの、そういったコンジョイントカードにはない要素が自身の家選びにおける重要項目であることも多いです。そのため、スコア自体がどの要素から来ているかいう予測というよりかは、今回勉強したコンジョイント分析では、カードにある各要素どうしを比較する、という目的で使うことを意識する必要がありそうですね（？）\n",
    "- コンジョイント分析では、コンジョイントカードの内容がカテゴリーデータなので、数値（例：家賃）が混じるとバリデーションが少なくなったりしてしまいますね。また、今回は等間隔に刻んだので問題なさそうですが、解析上は家賃の差分２万円ということは考慮せず、単にことなるカテゴリーデータとして扱うので、少し違和感がありました。また継続して勉強してみたいと思います。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f199d10b",
   "metadata": {},
   "source": [
    "# 参考記事  \n",
    "グロービス経営大学院:コンジョイント分析  \n",
    "https://mba.globis.ac.jp/about_mba/glossary/detail-11804.html  \n",
    "Pythonでコンジョイント分析に挑戦  \n",
    "https://wannko5296.hatenablog.com/entry/conjoint_analysis_in_python  "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.12"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": "20",
    "lenVar": "28"
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "position": {
    "height": "489.806px",
    "left": "1438.67px",
    "right": "20px",
    "top": "289px",
    "width": "350px"
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}