Day_028_HW.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# (Kaggle)房價預測\n",
    "以下用房價預測資料, 展示特徵篩選的作法"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 做完特征工程前的準備\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import os\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.linear_model import LinearRegression, LogisticRegression\n",
    "from sklearn.ensemble import GradientBoostingRegressor, GradientBoostingClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>MSSubClass</th>\n",
       "      <th>MSZoning</th>\n",
       "      <th>LotFrontage</th>\n",
       "      <th>LotArea</th>\n",
       "      <th>Street</th>\n",
       "      <th>Alley</th>\n",
       "      <th>LotShape</th>\n",
       "      <th>LandContour</th>\n",
       "      <th>Utilities</th>\n",
       "      <th>LotConfig</th>\n",
       "      <th>...</th>\n",
       "      <th>PoolArea</th>\n",
       "      <th>PoolQC</th>\n",
       "      <th>Fence</th>\n",
       "      <th>MiscFeature</th>\n",
       "      <th>MiscVal</th>\n",
       "      <th>MoSold</th>\n",
       "      <th>YrSold</th>\n",
       "      <th>SaleType</th>\n",
       "      <th>SaleCondition</th>\n",
       "      <th>SalePrice</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>60</td>\n",
       "      <td>RL</td>\n",
       "      <td>65.0</td>\n",
       "      <td>8450</td>\n",
       "      <td>Pave</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Reg</td>\n",
       "      <td>Lvl</td>\n",
       "      <td>AllPub</td>\n",
       "      <td>Inside</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2008</td>\n",
       "      <td>WD</td>\n",
       "      <td>Normal</td>\n",
       "      <td>208500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20</td>\n",
       "      <td>RL</td>\n",
       "      <td>80.0</td>\n",
       "      <td>9600</td>\n",
       "      <td>Pave</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Reg</td>\n",
       "      <td>Lvl</td>\n",
       "      <td>AllPub</td>\n",
       "      <td>FR2</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>2007</td>\n",
       "      <td>WD</td>\n",
       "      <td>Normal</td>\n",
       "      <td>181500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>60</td>\n",
       "      <td>RL</td>\n",
       "      <td>68.0</td>\n",
       "      <td>11250</td>\n",
       "      <td>Pave</td>\n",
       "      <td>NaN</td>\n",
       "      <td>IR1</td>\n",
       "      <td>Lvl</td>\n",
       "      <td>AllPub</td>\n",
       "      <td>Inside</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>2008</td>\n",
       "      <td>WD</td>\n",
       "      <td>Normal</td>\n",
       "      <td>223500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>70</td>\n",
       "      <td>RL</td>\n",
       "      <td>60.0</td>\n",
       "      <td>9550</td>\n",
       "      <td>Pave</td>\n",
       "      <td>NaN</td>\n",
       "      <td>IR1</td>\n",
       "      <td>Lvl</td>\n",
       "      <td>AllPub</td>\n",
       "      <td>Corner</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2006</td>\n",
       "      <td>WD</td>\n",
       "      <td>Abnorml</td>\n",
       "      <td>140000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>60</td>\n",
       "      <td>RL</td>\n",
       "      <td>84.0</td>\n",
       "      <td>14260</td>\n",
       "      <td>Pave</td>\n",
       "      <td>NaN</td>\n",
       "      <td>IR1</td>\n",
       "      <td>Lvl</td>\n",
       "      <td>AllPub</td>\n",
       "      <td>FR2</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>2008</td>\n",
       "      <td>WD</td>\n",
       "      <td>Normal</td>\n",
       "      <td>250000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 80 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   MSSubClass MSZoning  LotFrontage  LotArea Street Alley LotShape  \\\n",
       "0          60       RL         65.0     8450   Pave   NaN      Reg   \n",
       "1          20       RL         80.0     9600   Pave   NaN      Reg   \n",
       "2          60       RL         68.0    11250   Pave   NaN      IR1   \n",
       "3          70       RL         60.0     9550   Pave   NaN      IR1   \n",
       "4          60       RL         84.0    14260   Pave   NaN      IR1   \n",
       "\n",
       "  LandContour Utilities LotConfig    ...     PoolArea PoolQC Fence  \\\n",
       "0         Lvl    AllPub    Inside    ...            0    NaN   NaN   \n",
       "1         Lvl    AllPub       FR2    ...            0    NaN   NaN   \n",
       "2         Lvl    AllPub    Inside    ...            0    NaN   NaN   \n",
       "3         Lvl    AllPub    Corner    ...            0    NaN   NaN   \n",
       "4         Lvl    AllPub       FR2    ...            0    NaN   NaN   \n",
       "\n",
       "  MiscFeature MiscVal MoSold  YrSold  SaleType  SaleCondition  SalePrice  \n",
       "0         NaN       0      2    2008        WD         Normal     208500  \n",
       "1         NaN       0      5    2007        WD         Normal     181500  \n",
       "2         NaN       0      9    2008        WD         Normal     223500  \n",
       "3         NaN       0      2    2006        WD        Abnorml     140000  \n",
       "4         NaN       0     12    2008        WD         Normal     250000  \n",
       "\n",
       "[5 rows x 80 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = './data/'\n",
    "df = pd.read_csv(os.path.join(path,'house_train.csv'))\n",
    "train_Y = np.log1p(df['SalePrice'])\n",
    "df = df.drop(['Id'], axis=1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 計算df整體相關係數, 並繪製成熱圖\n",
    "sns.set_style(\"ticks\")\n",
    "corr = df.corr()\n",
    "cmap = sns.cubehelix_palette(start = 1, rot = 5, gamma=2, as_cmap = True) #matplotlib的colormap名稱或顏色對象\n",
    "sns.heatmap(corr,cmap=cmap,linewidths = 0.01) #linewidths區分隔線\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 刪除 SalePrice\n",
    "df = df.drop(['SalePrice'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 36 num features: \n",
      " ['MSSubClass', 'LotFrontage', 'LotArea', 'OverallQual', 'OverallCond', 'YearBuilt', 'YearRemodAdd', 'MasVnrArea', 'BsmtFinSF1', 'BsmtFinSF2', 'BsmtUnfSF', 'TotalBsmtSF', '1stFlrSF', '2ndFlrSF', 'LowQualFinSF', 'GrLivArea', 'BsmtFullBath', 'BsmtHalfBath', 'FullBath', 'HalfBath', 'BedroomAbvGr', 'KitchenAbvGr', 'TotRmsAbvGrd', 'Fireplaces', 'GarageYrBlt', 'GarageCars', 'GarageArea', 'WoodDeckSF', 'OpenPorchSF', 'EnclosedPorch', '3SsnPorch', 'ScreenPorch', 'PoolArea', 'MiscVal', 'MoSold', 'YrSold']\n",
      "\n",
      " df shape is : (1460, 36)\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>MSSubClass</th>\n",
       "      <th>LotFrontage</th>\n",
       "      <th>LotArea</th>\n",
       "      <th>OverallQual</th>\n",
       "      <th>OverallCond</th>\n",
       "      <th>YearBuilt</th>\n",
       "      <th>YearRemodAdd</th>\n",
       "      <th>MasVnrArea</th>\n",
       "      <th>BsmtFinSF1</th>\n",
       "      <th>BsmtFinSF2</th>\n",
       "      <th>...</th>\n",
       "      <th>GarageArea</th>\n",
       "      <th>WoodDeckSF</th>\n",
       "      <th>OpenPorchSF</th>\n",
       "      <th>EnclosedPorch</th>\n",
       "      <th>3SsnPorch</th>\n",
       "      <th>ScreenPorch</th>\n",
       "      <th>PoolArea</th>\n",
       "      <th>MiscVal</th>\n",
       "      <th>MoSold</th>\n",
       "      <th>YrSold</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>60</td>\n",
       "      <td>65.0</td>\n",
       "      <td>8450</td>\n",
       "      <td>7</td>\n",
       "      <td>5</td>\n",
       "      <td>2003</td>\n",
       "      <td>2003</td>\n",
       "      <td>196.0</td>\n",
       "      <td>706</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>548</td>\n",
       "      <td>0</td>\n",
       "      <td>61</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20</td>\n",
       "      <td>80.0</td>\n",
       "      <td>9600</td>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>1976</td>\n",
       "      <td>1976</td>\n",
       "      <td>0.0</td>\n",
       "      <td>978</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>460</td>\n",
       "      <td>298</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>2007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>60</td>\n",
       "      <td>68.0</td>\n",
       "      <td>11250</td>\n",
       "      <td>7</td>\n",
       "      <td>5</td>\n",
       "      <td>2001</td>\n",
       "      <td>2002</td>\n",
       "      <td>162.0</td>\n",
       "      <td>486</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>608</td>\n",
       "      <td>0</td>\n",
       "      <td>42</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>2008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>70</td>\n",
       "      <td>60.0</td>\n",
       "      <td>9550</td>\n",
       "      <td>7</td>\n",
       "      <td>5</td>\n",
       "      <td>1915</td>\n",
       "      <td>1970</td>\n",
       "      <td>0.0</td>\n",
       "      <td>216</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>642</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>272</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2006</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>60</td>\n",
       "      <td>84.0</td>\n",
       "      <td>14260</td>\n",
       "      <td>8</td>\n",
       "      <td>5</td>\n",
       "      <td>2000</td>\n",
       "      <td>2000</td>\n",
       "      <td>350.0</td>\n",
       "      <td>655</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>836</td>\n",
       "      <td>192</td>\n",
       "      <td>84</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>2008</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   MSSubClass  LotFrontage  LotArea  OverallQual  OverallCond  YearBuilt  \\\n",
       "0          60         65.0     8450            7            5       2003   \n",
       "1          20         80.0     9600            6            8       1976   \n",
       "2          60         68.0    11250            7            5       2001   \n",
       "3          70         60.0     9550            7            5       1915   \n",
       "4          60         84.0    14260            8            5       2000   \n",
       "\n",
       "   YearRemodAdd  MasVnrArea  BsmtFinSF1  BsmtFinSF2   ...    GarageArea  \\\n",
       "0          2003       196.0         706           0   ...           548   \n",
       "1          1976         0.0         978           0   ...           460   \n",
       "2          2002       162.0         486           0   ...           608   \n",
       "3          1970         0.0         216           0   ...           642   \n",
       "4          2000       350.0         655           0   ...           836   \n",
       "\n",
       "   WoodDeckSF  OpenPorchSF  EnclosedPorch  3SsnPorch  ScreenPorch  PoolArea  \\\n",
       "0           0           61              0          0            0         0   \n",
       "1         298            0              0          0            0         0   \n",
       "2           0           42              0          0            0         0   \n",
       "3           0           35            272          0            0         0   \n",
       "4         192           84              0          0            0         0   \n",
       "\n",
       "   MiscVal  MoSold  YrSold  \n",
       "0        0       2    2008  \n",
       "1        0       5    2007  \n",
       "2        0       9    2008  \n",
       "3        0       2    2006  \n",
       "4        0      12    2008  \n",
       "\n",
       "[5 rows x 36 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#只取 int64, float64 兩種數值型欄位, 存於 num_features 中\n",
    "num_features = []\n",
    "for dtype, feature in zip(df.dtypes, df.columns):\n",
    "    if dtype == 'int64' or dtype == 'float64':\n",
    "        num_features.append(feature)\n",
    "print(f\"There are {len(num_features)} num features: \\n {num_features}\")\n",
    "\n",
    "# 削減文字型欄位, 只剩數值型欄位\n",
    "df = df[num_features]\n",
    "df = df.fillna(-1)\n",
    "MMEncoder = MinMaxScaler()\n",
    "print(f'\\n df shape is : {df.shape}')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['LotFrontage', 'LotArea', 'OverallQual', 'YearBuilt', 'YearRemodAdd',\n",
      "       'MasVnrArea', 'BsmtFinSF1', 'BsmtUnfSF', 'TotalBsmtSF', '1stFlrSF',\n",
      "       '2ndFlrSF', 'GrLivArea', 'BsmtFullBath', 'FullBath', 'HalfBath',\n",
      "       'BedroomAbvGr', 'KitchenAbvGr', 'TotRmsAbvGrd', 'Fireplaces',\n",
      "       'GarageYrBlt', 'GarageCars', 'GarageArea', 'WoodDeckSF', 'OpenPorchSF',\n",
      "       'EnclosedPorch', 'ScreenPorch'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "# 篩選出相關係數大於 0.1 或小於 -0.1 的特徵\n",
    "high_list = corr['SalePrice'][abs(corr['SalePrice'])>0.1].index\n",
    "high_list = high_list[:-1]\n",
    "print(high_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8499683837163878"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 原始特徵 + 線性迴歸\n",
    "train_X = MMEncoder.fit_transform(df)\n",
    "estimator = LinearRegression()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8288821541594242"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 高相關性特徵 + 線性迴歸\n",
    "train_X = df[high_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "estimator = LinearRegression()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8859106154281248"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 原始特徵 + 梯度提升樹\n",
    "train_X = MMEncoder.fit_transform(df)\n",
    "estimator = GradientBoostingRegressor()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8743990370829569"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 高相關性特徵 + 梯度提升樹\n",
    "train_X = df[high_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "estimator = GradientBoostingRegressor()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([-0.12408194, -0.        ,  0.        ,  0.81411198,  0.2970946 ,\n",
       "        0.33007182,  0.08851589,  0.        ,  0.04287006,  0.        ,\n",
       "       -0.        ,  0.17513694,  0.11086056,  0.        , -0.        ,\n",
       "        0.87326328,  0.2040765 ,  0.        ,  0.10285172,  0.03174629,\n",
       "        0.        , -0.        ,  0.12246045,  0.17953898,  0.01851559,\n",
       "        0.27672243,  0.03230955,  0.09061659,  0.        ,  0.        ,\n",
       "        0.        ,  0.09550047, -0.        , -0.        ,  0.        ,\n",
       "       -0.01874705])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import Lasso\n",
    "L1_Reg = Lasso(alpha = 0.001)\n",
    "train_X = MMEncoder.fit_transform(df)\n",
    "L1_Reg.fit(train_X, train_Y)\n",
    "L1_Reg.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['MSSubClass',\n",
       " 'OverallQual',\n",
       " 'OverallCond',\n",
       " 'YearBuilt',\n",
       " 'YearRemodAdd',\n",
       " 'BsmtFinSF1',\n",
       " 'TotalBsmtSF',\n",
       " '1stFlrSF',\n",
       " 'GrLivArea',\n",
       " 'BsmtFullBath',\n",
       " 'FullBath',\n",
       " 'HalfBath',\n",
       " 'TotRmsAbvGrd',\n",
       " 'Fireplaces',\n",
       " 'GarageYrBlt',\n",
       " 'GarageCars',\n",
       " 'GarageArea',\n",
       " 'WoodDeckSF',\n",
       " 'ScreenPorch',\n",
       " 'YrSold']"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from itertools import compress\n",
    "L1_mask = list(abs(L1_Reg.coef_)>0)\n",
    "L1_list = list(compress(list(df),list(L1_mask)))\n",
    "L1_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8475901403725672"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# L1+Linear\n",
    "train_X = df[L1_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "estimator = LinearRegression()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8845950644643772"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# L1 + Gradient\n",
    "train_X = df[L1_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "estimator = GradientBoostingRegressor()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# 作業 : (Kaggle)鐵達尼生存預測\n",
    "***\n",
    "https://www.kaggle.com/c/titanic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>Survived</th>\n",
       "      <th>Pclass</th>\n",
       "      <th>Name</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Ticket</th>\n",
       "      <th>Fare</th>\n",
       "      <th>Cabin</th>\n",
       "      <th>Embarked</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>Braund, Mr. Owen Harris</td>\n",
       "      <td>male</td>\n",
       "      <td>22.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>A/5 21171</td>\n",
       "      <td>7.2500</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Cumings, Mrs. John Bradley (Florence Briggs Th...</td>\n",
       "      <td>female</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>PC 17599</td>\n",
       "      <td>71.2833</td>\n",
       "      <td>C85</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>Heikkinen, Miss. Laina</td>\n",
       "      <td>female</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>STON/O2. 3101282</td>\n",
       "      <td>7.9250</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Futrelle, Mrs. Jacques Heath (Lily May Peel)</td>\n",
       "      <td>female</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113803</td>\n",
       "      <td>53.1000</td>\n",
       "      <td>C123</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>Allen, Mr. William Henry</td>\n",
       "      <td>male</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>373450</td>\n",
       "      <td>8.0500</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Survived  Pclass                                               Name  \\\n",
       "0         0       3                            Braund, Mr. Owen Harris   \n",
       "1         1       1  Cumings, Mrs. John Bradley (Florence Briggs Th...   \n",
       "2         1       3                             Heikkinen, Miss. Laina   \n",
       "3         1       1       Futrelle, Mrs. Jacques Heath (Lily May Peel)   \n",
       "4         0       3                           Allen, Mr. William Henry   \n",
       "\n",
       "      Sex   Age  SibSp  Parch            Ticket     Fare Cabin Embarked  \n",
       "0    male  22.0      1      0         A/5 21171   7.2500   NaN        S  \n",
       "1  female  38.0      1      0          PC 17599  71.2833   C85        C  \n",
       "2  female  26.0      0      0  STON/O2. 3101282   7.9250   NaN        S  \n",
       "3  female  35.0      1      0            113803  53.1000  C123        S  \n",
       "4    male  35.0      0      0            373450   8.0500   NaN        S  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 做完特徵工程前的所有準備 (與前範例相同)\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import copy\n",
    "from sklearn.preprocessing import LabelEncoder, MinMaxScaler\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "data_path = 'data/'\n",
    "df = pd.read_csv(data_path + 'titanic_train.csv')\n",
    "\n",
    "train_Y = df['Survived']\n",
    "df = df.drop(['PassengerId'] , axis=1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 計算df整體相關係數, 並繪製成熱圖\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "corr = df.corr()\n",
    "sns.heatmap(corr)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5 Numeric Features : ['Pclass', 'Age', 'SibSp', 'Parch', 'Fare']\n",
      "\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>Pclass</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Fare</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3</td>\n",
       "      <td>22.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7.2500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>71.2833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.9250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>53.1000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0500</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Pclass   Age  SibSp  Parch     Fare\n",
       "0       3  22.0      1      0   7.2500\n",
       "1       1  38.0      1      0  71.2833\n",
       "2       3  26.0      0      0   7.9250\n",
       "3       1  35.0      1      0  53.1000\n",
       "4       3  35.0      0      0   8.0500"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 記得刪除 Survived\n",
    "df = df.drop(['Survived'] , axis=1)\n",
    "\n",
    "#只取 int64, float64 兩種數值型欄位, 存於 num_features 中\n",
    "num_features = []\n",
    "for dtype, feature in zip(df.dtypes, df.columns):\n",
    "    if dtype == 'float64' or dtype == 'int64':\n",
    "        num_features.append(feature)\n",
    "print(f'{len(num_features)} Numeric Features : {num_features}\\n')\n",
    "\n",
    "# 削減文字型欄位, 只剩數值型欄位\n",
    "df = df[num_features]\n",
    "df = df.fillna(-1)\n",
    "MMEncoder = MinMaxScaler()\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# 作業1\n",
    "* 鐵達尼生存率預測中，試著變更兩種以上的相關係數門檻值，觀察預測能力是否提升?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7038635542329971"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 原始特徵 + 邏輯斯迴歸\n",
    "train_X = MMEncoder.fit_transform(df)\n",
    "estimator = LogisticRegression()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['Pclass', 'Age', 'Parch', 'Fare'], dtype='object')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7049872206659582"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 篩選出相關係數大於 0.5 或小於 -0.5 的特徵\n",
    "high_list = corr['Survived'][abs(corr['Survived'])>0.05].index\n",
    "high_list = high_list[1:]\n",
    "print(high_list)\n",
    "# 特徵1 + 邏輯斯迴歸\n",
    "train_X = df[high_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['Pclass', 'Parch', 'Fare'], dtype='object')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.6836511764895717"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 篩選出相關係數大於 0.5 或小於 -0.5 的特徵\n",
    "high_list = corr['Survived'][abs(corr['Survived'])>0.08].index\n",
    "high_list = high_list[1:]\n",
    "print(high_list)\n",
    "# 特徵1 + 邏輯斯迴歸\n",
    "train_X = df[high_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 作業2\n",
    "* 續上題，使用 L1 Embedding 做特徵選擇(自訂門檻)，觀察預測能力是否提升?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([-0.4116012 , -0.18850803, -0.        ,  0.13666287,  0.01405623])"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import Lasso\n",
    "L1_Reg = Lasso(alpha = 0.003)\n",
    "\n",
    "train_X = MMEncoder.fit_transform(df)\n",
    "L1_Reg.fit(train_X, train_Y)\n",
    "L1_Reg.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Pclass', 'Age', 'Parch', 'Fare']"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from itertools import compress\n",
    "L1_mask = list(abs(L1_Reg.coef_)>0)\n",
    "L1_list = list(compress(list(df),list(L1_mask)))\n",
    "L1_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7049872206659582"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# L1 + Logist\n",
    "train_X = df[L1_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "estimator = LogisticRegression()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\TL_Yu\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:323: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7138811193895143"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# L1 + Gradient\n",
    "train_X = df[L1_list]\n",
    "train_X = MMEncoder.fit_transform(train_X)\n",
    "estimator = GradientBoostingClassifier()\n",
    "cross_val_score(estimator, train_X, train_Y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}