teach_Feature_Engineering（regression）.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ec09f09a",
   "metadata": {},
   "source": [
    "# 特征工程（回归预测）"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "42f1eb47",
   "metadata": {},
   "source": [
    "## 【任务描述】   \n",
    "特征工程实践操作:  \n",
    "1. 熟练掌握特征处理、特征选择的基本流程和方法。      \n",
    "2. 学会根据实际需求选择合适的特征处理方法。    \n",
    "3. 能够分析和解决特征处理过程中出现的\n",
    "    \n",
    "## 【任务准备】    \n",
    "1. 学员已完成回归数据预处理实践课程，掌握pandas、Matplotlib、Seaborn、sklearn等数据分析常用库的基本使用。    \n",
    "2. 准备好用于实践的数据集（本次使用回归数据预处理实践课程得到的已清洗数据）。    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e360a390",
   "metadata": {},
   "source": [
    "## 【任务实施】    \n",
    "### 步骤 1: 启动Jupyter Notebook    \n",
    "+ 在搜索栏输入\"cmd\"命令，启动命令提示符窗口。    \n",
    "+ 输入\"jupyter notebook\"命令，并按回车键启动Jupyter Notebook。   \n",
    "    \n",
    "### 步骤 2: 创建新的Notebook    \n",
    "+ 在Jupyter的Web界面中，点击右上角的 \"New\" 按钮。    \n",
    "+ 选择 \"Python 3\"内核来创建一个新的Python 3 Notebook。    \n",
    "    \n",
    "### 步骤 3: 导入必要的库\n",
    "+ 使用numpy和pandas库来进行数据清洗，使用sklearn库来进行数据的分类、离散化等特征处理和特征选择，使用scipy来进行统计分析，使用matplotlib和seaborn来进行数据的可视化。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "86dee898",
   "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 scipy.stats as st\n",
    "\n",
    "# 定义字体，在图表中正常显示汉字\n",
    "plt.rcParams['font.sans-serif'] = ['SimHei']\n",
    "# 在图表中正常显示负号\n",
    "plt.rcParams['axes.unicode_minus'] = False\n",
    "\n",
    "#导入warnings包，利用过滤器来实现忽略警告语句。\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67cb284f",
   "metadata": {},
   "source": [
    "### 步骤 4: 导入数据集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "83dc1737",
   "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>SaleID</th>\n",
       "      <th>name</th>\n",
       "      <th>regDate</th>\n",
       "      <th>model</th>\n",
       "      <th>brand</th>\n",
       "      <th>bodyType</th>\n",
       "      <th>fuelType</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>power</th>\n",
       "      <th>kilometer</th>\n",
       "      <th>...</th>\n",
       "      <th>v_5</th>\n",
       "      <th>v_6</th>\n",
       "      <th>v_7</th>\n",
       "      <th>v_8</th>\n",
       "      <th>v_9</th>\n",
       "      <th>v_10</th>\n",
       "      <th>v_11</th>\n",
       "      <th>v_12</th>\n",
       "      <th>v_13</th>\n",
       "      <th>v_14</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>736</td>\n",
       "      <td>20040402</td>\n",
       "      <td>30.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.235676</td>\n",
       "      <td>0.101988</td>\n",
       "      <td>0.129549</td>\n",
       "      <td>0.022816</td>\n",
       "      <td>0.097462</td>\n",
       "      <td>-2.881803</td>\n",
       "      <td>2.804097</td>\n",
       "      <td>-2.420821</td>\n",
       "      <td>0.795292</td>\n",
       "      <td>0.914762</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2262</td>\n",
       "      <td>20030301</td>\n",
       "      <td>40.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.264777</td>\n",
       "      <td>0.121004</td>\n",
       "      <td>0.135731</td>\n",
       "      <td>0.026597</td>\n",
       "      <td>0.020582</td>\n",
       "      <td>-4.900482</td>\n",
       "      <td>2.096338</td>\n",
       "      <td>-1.030483</td>\n",
       "      <td>-1.722674</td>\n",
       "      <td>0.245522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>14874</td>\n",
       "      <td>20040403</td>\n",
       "      <td>115.0</td>\n",
       "      <td>15</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>163.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.251410</td>\n",
       "      <td>0.114912</td>\n",
       "      <td>0.165147</td>\n",
       "      <td>0.062173</td>\n",
       "      <td>0.027075</td>\n",
       "      <td>-4.846749</td>\n",
       "      <td>1.803559</td>\n",
       "      <td>1.565330</td>\n",
       "      <td>-0.832687</td>\n",
       "      <td>-0.229963</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>71865</td>\n",
       "      <td>19960908</td>\n",
       "      <td>109.0</td>\n",
       "      <td>10</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>193.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.274293</td>\n",
       "      <td>0.110300</td>\n",
       "      <td>0.121964</td>\n",
       "      <td>0.033395</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-4.509599</td>\n",
       "      <td>1.285940</td>\n",
       "      <td>-0.501868</td>\n",
       "      <td>-2.438353</td>\n",
       "      <td>-0.478699</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>111080</td>\n",
       "      <td>20120103</td>\n",
       "      <td>110.0</td>\n",
       "      <td>5</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.228036</td>\n",
       "      <td>0.073205</td>\n",
       "      <td>0.091880</td>\n",
       "      <td>0.078819</td>\n",
       "      <td>0.121534</td>\n",
       "      <td>-1.896240</td>\n",
       "      <td>0.910783</td>\n",
       "      <td>0.931110</td>\n",
       "      <td>2.834518</td>\n",
       "      <td>1.923482</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149995</th>\n",
       "      <td>149995</td>\n",
       "      <td>163978</td>\n",
       "      <td>20000607</td>\n",
       "      <td>121.0</td>\n",
       "      <td>10</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>163.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.280264</td>\n",
       "      <td>0.000310</td>\n",
       "      <td>0.048441</td>\n",
       "      <td>0.071158</td>\n",
       "      <td>0.019174</td>\n",
       "      <td>1.988114</td>\n",
       "      <td>-2.983973</td>\n",
       "      <td>0.589167</td>\n",
       "      <td>-1.304370</td>\n",
       "      <td>-0.302592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149996</th>\n",
       "      <td>149996</td>\n",
       "      <td>184535</td>\n",
       "      <td>20091102</td>\n",
       "      <td>116.0</td>\n",
       "      <td>11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>125.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.253217</td>\n",
       "      <td>0.000777</td>\n",
       "      <td>0.084079</td>\n",
       "      <td>0.099681</td>\n",
       "      <td>0.079371</td>\n",
       "      <td>1.839166</td>\n",
       "      <td>-2.774615</td>\n",
       "      <td>2.553994</td>\n",
       "      <td>0.924196</td>\n",
       "      <td>-0.272160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149997</th>\n",
       "      <td>149997</td>\n",
       "      <td>147587</td>\n",
       "      <td>20101003</td>\n",
       "      <td>60.0</td>\n",
       "      <td>11</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.233353</td>\n",
       "      <td>0.000705</td>\n",
       "      <td>0.118872</td>\n",
       "      <td>0.100118</td>\n",
       "      <td>0.097914</td>\n",
       "      <td>2.439812</td>\n",
       "      <td>-1.630677</td>\n",
       "      <td>2.290197</td>\n",
       "      <td>1.891922</td>\n",
       "      <td>0.414931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149998</th>\n",
       "      <td>149998</td>\n",
       "      <td>45907</td>\n",
       "      <td>20060312</td>\n",
       "      <td>34.0</td>\n",
       "      <td>10</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>156.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.256369</td>\n",
       "      <td>0.000252</td>\n",
       "      <td>0.081479</td>\n",
       "      <td>0.083558</td>\n",
       "      <td>0.081498</td>\n",
       "      <td>2.075380</td>\n",
       "      <td>-2.633719</td>\n",
       "      <td>1.414937</td>\n",
       "      <td>0.431981</td>\n",
       "      <td>-1.659014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149999</th>\n",
       "      <td>149999</td>\n",
       "      <td>177672</td>\n",
       "      <td>19990204</td>\n",
       "      <td>19.0</td>\n",
       "      <td>28</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>193.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.284475</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.040072</td>\n",
       "      <td>0.062543</td>\n",
       "      <td>0.025819</td>\n",
       "      <td>1.978453</td>\n",
       "      <td>-3.179913</td>\n",
       "      <td>0.031724</td>\n",
       "      <td>-1.483350</td>\n",
       "      <td>-0.342674</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>150000 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        SaleID    name   regDate  model  brand  bodyType  fuelType  gearbox  \\\n",
       "0            0     736  20040402   30.0      6       1.0       0.0      0.0   \n",
       "1            1    2262  20030301   40.0      1       2.0       0.0      0.0   \n",
       "2            2   14874  20040403  115.0     15       1.0       0.0      0.0   \n",
       "3            3   71865  19960908  109.0     10       0.0       0.0      1.0   \n",
       "4            4  111080  20120103  110.0      5       1.0       0.0      0.0   \n",
       "...        ...     ...       ...    ...    ...       ...       ...      ...   \n",
       "149995  149995  163978  20000607  121.0     10       4.0       0.0      1.0   \n",
       "149996  149996  184535  20091102  116.0     11       0.0       0.0      0.0   \n",
       "149997  149997  147587  20101003   60.0     11       1.0       1.0      0.0   \n",
       "149998  149998   45907  20060312   34.0     10       3.0       1.0      0.0   \n",
       "149999  149999  177672  19990204   19.0     28       6.0       0.0      1.0   \n",
       "\n",
       "        power  kilometer  ...       v_5       v_6       v_7       v_8  \\\n",
       "0        60.0       12.5  ...  0.235676  0.101988  0.129549  0.022816   \n",
       "1         0.0       15.0  ...  0.264777  0.121004  0.135731  0.026597   \n",
       "2       163.0       12.5  ...  0.251410  0.114912  0.165147  0.062173   \n",
       "3       193.0       15.0  ...  0.274293  0.110300  0.121964  0.033395   \n",
       "4        68.0        5.0  ...  0.228036  0.073205  0.091880  0.078819   \n",
       "...       ...        ...  ...       ...       ...       ...       ...   \n",
       "149995  163.0       15.0  ...  0.280264  0.000310  0.048441  0.071158   \n",
       "149996  125.0       10.0  ...  0.253217  0.000777  0.084079  0.099681   \n",
       "149997   90.0        6.0  ...  0.233353  0.000705  0.118872  0.100118   \n",
       "149998  156.0       15.0  ...  0.256369  0.000252  0.081479  0.083558   \n",
       "149999  193.0       12.5  ...  0.284475  0.000000  0.040072  0.062543   \n",
       "\n",
       "             v_9      v_10      v_11      v_12      v_13      v_14  \n",
       "0       0.097462 -2.881803  2.804097 -2.420821  0.795292  0.914762  \n",
       "1       0.020582 -4.900482  2.096338 -1.030483 -1.722674  0.245522  \n",
       "2       0.027075 -4.846749  1.803559  1.565330 -0.832687 -0.229963  \n",
       "3       0.000000 -4.509599  1.285940 -0.501868 -2.438353 -0.478699  \n",
       "4       0.121534 -1.896240  0.910783  0.931110  2.834518  1.923482  \n",
       "...          ...       ...       ...       ...       ...       ...  \n",
       "149995  0.019174  1.988114 -2.983973  0.589167 -1.304370 -0.302592  \n",
       "149996  0.079371  1.839166 -2.774615  2.553994  0.924196 -0.272160  \n",
       "149997  0.097914  2.439812 -1.630677  2.290197  1.891922  0.414931  \n",
       "149998  0.081498  2.075380 -2.633719  1.414937  0.431981 -1.659014  \n",
       "149999  0.025819  1.978453 -3.179913  0.031724 -1.483350 -0.342674  \n",
       "\n",
       "[150000 rows x 31 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 用Pandas导入CSV文件\n",
    "train = pd.read_csv('used_car_train_clean.csv')\n",
    "test = pd.read_csv('used_car_test_clean.csv')\n",
    "train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "205e4343",
   "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>SaleID</th>\n",
       "      <th>name</th>\n",
       "      <th>regDate</th>\n",
       "      <th>model</th>\n",
       "      <th>brand</th>\n",
       "      <th>bodyType</th>\n",
       "      <th>fuelType</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>power</th>\n",
       "      <th>kilometer</th>\n",
       "      <th>...</th>\n",
       "      <th>v_5</th>\n",
       "      <th>v_6</th>\n",
       "      <th>v_7</th>\n",
       "      <th>v_8</th>\n",
       "      <th>v_9</th>\n",
       "      <th>v_10</th>\n",
       "      <th>v_11</th>\n",
       "      <th>v_12</th>\n",
       "      <th>v_13</th>\n",
       "      <th>v_14</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>150000</td>\n",
       "      <td>66932</td>\n",
       "      <td>20111212</td>\n",
       "      <td>222.0</td>\n",
       "      <td>4</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>313</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.264405</td>\n",
       "      <td>0.121800</td>\n",
       "      <td>0.070899</td>\n",
       "      <td>0.106558</td>\n",
       "      <td>0.078867</td>\n",
       "      <td>-7.050969</td>\n",
       "      <td>-0.854626</td>\n",
       "      <td>4.800151</td>\n",
       "      <td>0.620011</td>\n",
       "      <td>-3.664654</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>150001</td>\n",
       "      <td>174960</td>\n",
       "      <td>19990211</td>\n",
       "      <td>19.0</td>\n",
       "      <td>21</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>75</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.261745</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.096733</td>\n",
       "      <td>0.013705</td>\n",
       "      <td>0.052383</td>\n",
       "      <td>3.679418</td>\n",
       "      <td>-0.729039</td>\n",
       "      <td>-3.796107</td>\n",
       "      <td>-1.541230</td>\n",
       "      <td>-0.757055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>150002</td>\n",
       "      <td>5356</td>\n",
       "      <td>20090304</td>\n",
       "      <td>82.0</td>\n",
       "      <td>21</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>109</td>\n",
       "      <td>7.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.260216</td>\n",
       "      <td>0.112081</td>\n",
       "      <td>0.078082</td>\n",
       "      <td>0.062078</td>\n",
       "      <td>0.050540</td>\n",
       "      <td>-4.926690</td>\n",
       "      <td>1.001106</td>\n",
       "      <td>0.826562</td>\n",
       "      <td>0.138226</td>\n",
       "      <td>0.754033</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>150003</td>\n",
       "      <td>50688</td>\n",
       "      <td>20100405</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>160</td>\n",
       "      <td>7.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.260466</td>\n",
       "      <td>0.106727</td>\n",
       "      <td>0.081146</td>\n",
       "      <td>0.075971</td>\n",
       "      <td>0.048268</td>\n",
       "      <td>-4.864637</td>\n",
       "      <td>0.505493</td>\n",
       "      <td>1.870379</td>\n",
       "      <td>0.366038</td>\n",
       "      <td>1.312775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>150004</td>\n",
       "      <td>161428</td>\n",
       "      <td>19970703</td>\n",
       "      <td>26.0</td>\n",
       "      <td>14</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>75</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.250999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.077806</td>\n",
       "      <td>0.028600</td>\n",
       "      <td>0.081709</td>\n",
       "      <td>3.616475</td>\n",
       "      <td>-0.673236</td>\n",
       "      <td>-3.197685</td>\n",
       "      <td>-0.025678</td>\n",
       "      <td>-0.101290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49995</th>\n",
       "      <td>199995</td>\n",
       "      <td>20903</td>\n",
       "      <td>19960503</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>116</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.284664</td>\n",
       "      <td>0.130044</td>\n",
       "      <td>0.049833</td>\n",
       "      <td>0.028807</td>\n",
       "      <td>0.004616</td>\n",
       "      <td>-5.978511</td>\n",
       "      <td>1.303174</td>\n",
       "      <td>-1.207191</td>\n",
       "      <td>-1.981240</td>\n",
       "      <td>-0.357695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49996</th>\n",
       "      <td>199996</td>\n",
       "      <td>708</td>\n",
       "      <td>19991011</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>75</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.268101</td>\n",
       "      <td>0.108095</td>\n",
       "      <td>0.066039</td>\n",
       "      <td>0.025468</td>\n",
       "      <td>0.025971</td>\n",
       "      <td>-3.913825</td>\n",
       "      <td>1.759524</td>\n",
       "      <td>-2.075658</td>\n",
       "      <td>-1.154847</td>\n",
       "      <td>0.169073</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49997</th>\n",
       "      <td>199997</td>\n",
       "      <td>6693</td>\n",
       "      <td>20040412</td>\n",
       "      <td>49.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>224</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.269432</td>\n",
       "      <td>0.105724</td>\n",
       "      <td>0.117652</td>\n",
       "      <td>0.057479</td>\n",
       "      <td>0.015669</td>\n",
       "      <td>-4.639065</td>\n",
       "      <td>0.654713</td>\n",
       "      <td>1.137756</td>\n",
       "      <td>-1.390531</td>\n",
       "      <td>0.254420</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49998</th>\n",
       "      <td>199998</td>\n",
       "      <td>96900</td>\n",
       "      <td>20020008</td>\n",
       "      <td>27.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>334</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.261152</td>\n",
       "      <td>0.000490</td>\n",
       "      <td>0.137366</td>\n",
       "      <td>0.086216</td>\n",
       "      <td>0.051383</td>\n",
       "      <td>1.833504</td>\n",
       "      <td>-2.828687</td>\n",
       "      <td>2.465630</td>\n",
       "      <td>-0.911682</td>\n",
       "      <td>-2.057353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49999</th>\n",
       "      <td>199999</td>\n",
       "      <td>193384</td>\n",
       "      <td>20041109</td>\n",
       "      <td>166.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>68</td>\n",
       "      <td>9.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.228730</td>\n",
       "      <td>0.000300</td>\n",
       "      <td>0.103534</td>\n",
       "      <td>0.080625</td>\n",
       "      <td>0.124264</td>\n",
       "      <td>2.914571</td>\n",
       "      <td>-1.135270</td>\n",
       "      <td>0.547628</td>\n",
       "      <td>2.094057</td>\n",
       "      <td>-1.552150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>50000 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       SaleID    name   regDate  model  brand  bodyType  fuelType  gearbox  \\\n",
       "0      150000   66932  20111212  222.0      4       5.0       1.0      1.0   \n",
       "1      150001  174960  19990211   19.0     21       0.0       0.0      0.0   \n",
       "2      150002    5356  20090304   82.0     21       0.0       0.0      0.0   \n",
       "3      150003   50688  20100405    0.0      0       0.0       0.0      1.0   \n",
       "4      150004  161428  19970703   26.0     14       2.0       0.0      0.0   \n",
       "...       ...     ...       ...    ...    ...       ...       ...      ...   \n",
       "49995  199995   20903  19960503    4.0      4       4.0       0.0      0.0   \n",
       "49996  199996     708  19991011    0.0      0       0.0       0.0      0.0   \n",
       "49997  199997    6693  20040412   49.0      1       0.0       1.0      1.0   \n",
       "49998  199998   96900  20020008   27.0      1       0.0       0.0      1.0   \n",
       "49999  199999  193384  20041109  166.0      6       1.0       NaN      1.0   \n",
       "\n",
       "       power  kilometer  ...       v_5       v_6       v_7       v_8  \\\n",
       "0        313       15.0  ...  0.264405  0.121800  0.070899  0.106558   \n",
       "1         75       12.5  ...  0.261745  0.000000  0.096733  0.013705   \n",
       "2        109        7.0  ...  0.260216  0.112081  0.078082  0.062078   \n",
       "3        160        7.0  ...  0.260466  0.106727  0.081146  0.075971   \n",
       "4         75       15.0  ...  0.250999  0.000000  0.077806  0.028600   \n",
       "...      ...        ...  ...       ...       ...       ...       ...   \n",
       "49995    116       15.0  ...  0.284664  0.130044  0.049833  0.028807   \n",
       "49996     75       15.0  ...  0.268101  0.108095  0.066039  0.025468   \n",
       "49997    224       15.0  ...  0.269432  0.105724  0.117652  0.057479   \n",
       "49998    334       15.0  ...  0.261152  0.000490  0.137366  0.086216   \n",
       "49999     68        9.0  ...  0.228730  0.000300  0.103534  0.080625   \n",
       "\n",
       "            v_9      v_10      v_11      v_12      v_13      v_14  \n",
       "0      0.078867 -7.050969 -0.854626  4.800151  0.620011 -3.664654  \n",
       "1      0.052383  3.679418 -0.729039 -3.796107 -1.541230 -0.757055  \n",
       "2      0.050540 -4.926690  1.001106  0.826562  0.138226  0.754033  \n",
       "3      0.048268 -4.864637  0.505493  1.870379  0.366038  1.312775  \n",
       "4      0.081709  3.616475 -0.673236 -3.197685 -0.025678 -0.101290  \n",
       "...         ...       ...       ...       ...       ...       ...  \n",
       "49995  0.004616 -5.978511  1.303174 -1.207191 -1.981240 -0.357695  \n",
       "49996  0.025971 -3.913825  1.759524 -2.075658 -1.154847  0.169073  \n",
       "49997  0.015669 -4.639065  0.654713  1.137756 -1.390531  0.254420  \n",
       "49998  0.051383  1.833504 -2.828687  2.465630 -0.911682 -2.057353  \n",
       "49999  0.124264  2.914571 -1.135270  0.547628  2.094057 -1.552150  \n",
       "\n",
       "[50000 rows x 30 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e395325",
   "metadata": {},
   "source": [
    "### 步骤 5: 合并训练集和测试集 \n",
    "严格从统计学角度来说，合并训练集和测试集一起做特征，会导致训练数据获得一部分测试数据的信息，从而造成一定程度的数据泄露。    \n",
    "但是，训练集和测试集数据分布相差不大的情况下，这个数据泄露通常并不严重，造成的影响相对较小。    \n",
    "在数据竞赛中，将训练集和验证集合并来做特征，是一种常见的做法，除非特别声明。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6a22c899",
   "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>SaleID</th>\n",
       "      <th>name</th>\n",
       "      <th>regDate</th>\n",
       "      <th>model</th>\n",
       "      <th>brand</th>\n",
       "      <th>bodyType</th>\n",
       "      <th>fuelType</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>power</th>\n",
       "      <th>kilometer</th>\n",
       "      <th>...</th>\n",
       "      <th>v_6</th>\n",
       "      <th>v_7</th>\n",
       "      <th>v_8</th>\n",
       "      <th>v_9</th>\n",
       "      <th>v_10</th>\n",
       "      <th>v_11</th>\n",
       "      <th>v_12</th>\n",
       "      <th>v_13</th>\n",
       "      <th>v_14</th>\n",
       "      <th>train</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>736</td>\n",
       "      <td>20040402</td>\n",
       "      <td>30.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.101988</td>\n",
       "      <td>0.129549</td>\n",
       "      <td>0.022816</td>\n",
       "      <td>0.097462</td>\n",
       "      <td>-2.881803</td>\n",
       "      <td>2.804097</td>\n",
       "      <td>-2.420821</td>\n",
       "      <td>0.795292</td>\n",
       "      <td>0.914762</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2262</td>\n",
       "      <td>20030301</td>\n",
       "      <td>40.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.121004</td>\n",
       "      <td>0.135731</td>\n",
       "      <td>0.026597</td>\n",
       "      <td>0.020582</td>\n",
       "      <td>-4.900482</td>\n",
       "      <td>2.096338</td>\n",
       "      <td>-1.030483</td>\n",
       "      <td>-1.722674</td>\n",
       "      <td>0.245522</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>14874</td>\n",
       "      <td>20040403</td>\n",
       "      <td>115.0</td>\n",
       "      <td>15</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>163.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.114912</td>\n",
       "      <td>0.165147</td>\n",
       "      <td>0.062173</td>\n",
       "      <td>0.027075</td>\n",
       "      <td>-4.846749</td>\n",
       "      <td>1.803559</td>\n",
       "      <td>1.565330</td>\n",
       "      <td>-0.832687</td>\n",
       "      <td>-0.229963</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>71865</td>\n",
       "      <td>19960908</td>\n",
       "      <td>109.0</td>\n",
       "      <td>10</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>193.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.110300</td>\n",
       "      <td>0.121964</td>\n",
       "      <td>0.033395</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-4.509599</td>\n",
       "      <td>1.285940</td>\n",
       "      <td>-0.501868</td>\n",
       "      <td>-2.438353</td>\n",
       "      <td>-0.478699</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>111080</td>\n",
       "      <td>20120103</td>\n",
       "      <td>110.0</td>\n",
       "      <td>5</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.073205</td>\n",
       "      <td>0.091880</td>\n",
       "      <td>0.078819</td>\n",
       "      <td>0.121534</td>\n",
       "      <td>-1.896240</td>\n",
       "      <td>0.910783</td>\n",
       "      <td>0.931110</td>\n",
       "      <td>2.834518</td>\n",
       "      <td>1.923482</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199995</th>\n",
       "      <td>199995</td>\n",
       "      <td>20903</td>\n",
       "      <td>19960503</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>116.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.130044</td>\n",
       "      <td>0.049833</td>\n",
       "      <td>0.028807</td>\n",
       "      <td>0.004616</td>\n",
       "      <td>-5.978511</td>\n",
       "      <td>1.303174</td>\n",
       "      <td>-1.207191</td>\n",
       "      <td>-1.981240</td>\n",
       "      <td>-0.357695</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199996</th>\n",
       "      <td>199996</td>\n",
       "      <td>708</td>\n",
       "      <td>19991011</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.108095</td>\n",
       "      <td>0.066039</td>\n",
       "      <td>0.025468</td>\n",
       "      <td>0.025971</td>\n",
       "      <td>-3.913825</td>\n",
       "      <td>1.759524</td>\n",
       "      <td>-2.075658</td>\n",
       "      <td>-1.154847</td>\n",
       "      <td>0.169073</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199997</th>\n",
       "      <td>199997</td>\n",
       "      <td>6693</td>\n",
       "      <td>20040412</td>\n",
       "      <td>49.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>224.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.105724</td>\n",
       "      <td>0.117652</td>\n",
       "      <td>0.057479</td>\n",
       "      <td>0.015669</td>\n",
       "      <td>-4.639065</td>\n",
       "      <td>0.654713</td>\n",
       "      <td>1.137756</td>\n",
       "      <td>-1.390531</td>\n",
       "      <td>0.254420</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199998</th>\n",
       "      <td>199998</td>\n",
       "      <td>96900</td>\n",
       "      <td>20020008</td>\n",
       "      <td>27.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>334.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000490</td>\n",
       "      <td>0.137366</td>\n",
       "      <td>0.086216</td>\n",
       "      <td>0.051383</td>\n",
       "      <td>1.833504</td>\n",
       "      <td>-2.828687</td>\n",
       "      <td>2.465630</td>\n",
       "      <td>-0.911682</td>\n",
       "      <td>-2.057353</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199999</th>\n",
       "      <td>199999</td>\n",
       "      <td>193384</td>\n",
       "      <td>20041109</td>\n",
       "      <td>166.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000300</td>\n",
       "      <td>0.103534</td>\n",
       "      <td>0.080625</td>\n",
       "      <td>0.124264</td>\n",
       "      <td>2.914571</td>\n",
       "      <td>-1.135270</td>\n",
       "      <td>0.547628</td>\n",
       "      <td>2.094057</td>\n",
       "      <td>-1.552150</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>200000 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        SaleID    name   regDate  model  brand  bodyType  fuelType  gearbox  \\\n",
       "0            0     736  20040402   30.0      6       1.0       0.0      0.0   \n",
       "1            1    2262  20030301   40.0      1       2.0       0.0      0.0   \n",
       "2            2   14874  20040403  115.0     15       1.0       0.0      0.0   \n",
       "3            3   71865  19960908  109.0     10       0.0       0.0      1.0   \n",
       "4            4  111080  20120103  110.0      5       1.0       0.0      0.0   \n",
       "...        ...     ...       ...    ...    ...       ...       ...      ...   \n",
       "199995  199995   20903  19960503    4.0      4       4.0       0.0      0.0   \n",
       "199996  199996     708  19991011    0.0      0       0.0       0.0      0.0   \n",
       "199997  199997    6693  20040412   49.0      1       0.0       1.0      1.0   \n",
       "199998  199998   96900  20020008   27.0      1       0.0       0.0      1.0   \n",
       "199999  199999  193384  20041109  166.0      6       1.0       NaN      1.0   \n",
       "\n",
       "        power  kilometer  ...       v_6       v_7       v_8       v_9  \\\n",
       "0        60.0       12.5  ...  0.101988  0.129549  0.022816  0.097462   \n",
       "1         0.0       15.0  ...  0.121004  0.135731  0.026597  0.020582   \n",
       "2       163.0       12.5  ...  0.114912  0.165147  0.062173  0.027075   \n",
       "3       193.0       15.0  ...  0.110300  0.121964  0.033395  0.000000   \n",
       "4        68.0        5.0  ...  0.073205  0.091880  0.078819  0.121534   \n",
       "...       ...        ...  ...       ...       ...       ...       ...   \n",
       "199995  116.0       15.0  ...  0.130044  0.049833  0.028807  0.004616   \n",
       "199996   75.0       15.0  ...  0.108095  0.066039  0.025468  0.025971   \n",
       "199997  224.0       15.0  ...  0.105724  0.117652  0.057479  0.015669   \n",
       "199998  334.0       15.0  ...  0.000490  0.137366  0.086216  0.051383   \n",
       "199999   68.0        9.0  ...  0.000300  0.103534  0.080625  0.124264   \n",
       "\n",
       "            v_10      v_11      v_12      v_13      v_14  train  \n",
       "0      -2.881803  2.804097 -2.420821  0.795292  0.914762      1  \n",
       "1      -4.900482  2.096338 -1.030483 -1.722674  0.245522      1  \n",
       "2      -4.846749  1.803559  1.565330 -0.832687 -0.229963      1  \n",
       "3      -4.509599  1.285940 -0.501868 -2.438353 -0.478699      1  \n",
       "4      -1.896240  0.910783  0.931110  2.834518  1.923482      1  \n",
       "...          ...       ...       ...       ...       ...    ...  \n",
       "199995 -5.978511  1.303174 -1.207191 -1.981240 -0.357695      0  \n",
       "199996 -3.913825  1.759524 -2.075658 -1.154847  0.169073      0  \n",
       "199997 -4.639065  0.654713  1.137756 -1.390531  0.254420      0  \n",
       "199998  1.833504 -2.828687  2.465630 -0.911682 -2.057353      0  \n",
       "199999  2.914571 -1.135270  0.547628  2.094057 -1.552150      0  \n",
       "\n",
       "[200000 rows x 32 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 训练集和测试集放在一起，方便构造特征\n",
    "train['train']=1\n",
    "test['train']=0\n",
    "data = pd.concat([train, test], ignore_index=True, sort=False)\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "df8aec0a",
   "metadata": {},
   "source": [
    "### 步骤 6: 特征相关性分析"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "0c4ca50d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "price                1.000000\n",
      "v_12                 0.692823\n",
      "v_8                  0.685798\n",
      "v_0                  0.628397\n",
      "regDate              0.611959\n",
      "power                0.509806\n",
      "gearbox              0.329075\n",
      "bodyType             0.241303\n",
      "fuelType             0.200536\n",
      "v_5                  0.164317\n",
      "model                0.136983\n",
      "v_2                  0.085322\n",
      "v_6                  0.068970\n",
      "v_1                  0.060914\n",
      "v_14                 0.035911\n",
      "regionCode           0.014036\n",
      "creatDate            0.002955\n",
      "name                 0.002030\n",
      "SaleID              -0.001043\n",
      "seller              -0.002004\n",
      "v_13                -0.013993\n",
      "brand               -0.043799\n",
      "v_7                 -0.053024\n",
      "v_4                 -0.147085\n",
      "notRepairedDamage   -0.190623\n",
      "v_9                 -0.206205\n",
      "v_10                -0.246175\n",
      "v_11                -0.275320\n",
      "kilometer           -0.440519\n",
      "v_3                 -0.730946\n",
      "offerType                 NaN\n",
      "train                     NaN\n",
      "Name: price, dtype: float64 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "correlation = data.corr()\n",
    "print(correlation['price'].sort_values(ascending = False),'\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "edb30fb9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 300x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(3,10))\n",
    "sns.heatmap(data.corr().loc[:,['price']], annot=True)\n",
    "plt.title('各特征与价格相关性热力图')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "57c82a72",
   "metadata": {},
   "source": [
    "### 步骤 7: 构造时间特征    \n",
    "构造“使用时间”特征：data['creatDate'] - data['regDate']，反应汽车使用时间，一般来说价格与使用时间成反比。\n",
    "不过要注意，数据里有时间出错的格式，所以我们需要 errors='coerce'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "96f1fdcf",
   "metadata": {},
   "outputs": [],
   "source": [
    "data['used_time'] = (pd.to_datetime(data['creatDate'], format='%Y%m%d', errors='coerce') - \n",
    "                            pd.to_datetime(data['regDate'], format='%Y%m%d', errors='coerce')).dt.days"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1f98c60",
   "metadata": {},
   "source": [
    "看一下空数据，有一万多个样本的时间是有问题的，我们可以选择删除，也可以选择放着。    \n",
    "但是这里不建议删除，因为删除缺失数据占总样本量过大，7.5%。    \n",
    "我们可以先放着，如果使用树模型来预测，其本身就能处理缺失值，所以可以不用管。    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "78fed6d9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "15101"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['used_time'].isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "122067a6",
   "metadata": {},
   "source": [
    "### 步骤 8: 利用业务知识提取地理信息    \n",
    "因为是德国的数据，所以参考德国的邮编，相当于加入了先验知识，可以从邮编中提取城市信息。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "980eff22",
   "metadata": {},
   "outputs": [],
   "source": [
    "data['city'] = data['regionCode'].apply(lambda x : str(x)[:-3])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9458bf57",
   "metadata": {},
   "source": [
    "### 步骤 9: 查看特征的偏度和峰值      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "6d00bd1e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SaleID          Skewness: 00.00     Kurtosis: -01.20\n",
      "name            Skewness: 00.56     Kurtosis: -01.04\n",
      "regDate         Skewness: 00.03     Kurtosis: -00.69\n",
      "model           Skewness: 01.49     Kurtosis: 001.75\n",
      "brand           Skewness: 01.15     Kurtosis: 001.07\n",
      "bodyType        Skewness: 00.99     Kurtosis: 000.21\n",
      "fuelType        Skewness: 01.59     Kurtosis: 005.85\n",
      "gearbox         Skewness: 01.32     Kurtosis: -00.26\n",
      "power           Skewness: 84.74     Kurtosis: 12456.16\n",
      "kilometer       Skewness: -1.52     Kurtosis: 001.14\n",
      "notRepairedDamage Skewness: 02.44     Kurtosis: 003.94\n",
      "regionCode      Skewness: 00.68     Kurtosis: -00.35\n",
      "seller          Skewness: 447.21     Kurtosis: 200000.00\n",
      "offerType       Skewness: 00.00     Kurtosis: 000.00\n",
      "creatDate       Skewness: -82.93     Kurtosis: 7675.13\n",
      "v_0             Skewness: -1.30     Kurtosis: 003.97\n",
      "v_1             Skewness: 00.36     Kurtosis: -01.75\n",
      "v_2             Skewness: 04.87     Kurtosis: 024.21\n",
      "v_3             Skewness: 00.11     Kurtosis: -00.41\n",
      "v_4             Skewness: 00.37     Kurtosis: -00.20\n",
      "v_5             Skewness: -4.77     Kurtosis: 023.27\n",
      "v_6             Skewness: 00.37     Kurtosis: -01.74\n",
      "v_7             Skewness: 05.17     Kurtosis: 026.27\n",
      "v_8             Skewness: 00.20     Kurtosis: -00.64\n",
      "v_9             Skewness: 00.42     Kurtosis: -00.32\n",
      "v_10            Skewness: 00.02     Kurtosis: -00.58\n",
      "v_11            Skewness: 03.03     Kurtosis: 012.65\n",
      "v_12            Skewness: 00.36     Kurtosis: 000.26\n",
      "v_13            Skewness: 00.27     Kurtosis: -00.44\n",
      "v_14            Skewness: -1.18     Kurtosis: 002.37\n",
      "train           Skewness: -1.15     Kurtosis: -00.67\n",
      "used_time       Skewness: 00.03     Kurtosis: -00.69\n"
     ]
    }
   ],
   "source": [
    "# 删去预测值\n",
    "numeric_features = list(data.columns)\n",
    "numeric_features.remove('price')\n",
    "# 循环查看每一个特征\n",
    "for col in numeric_features:\n",
    "    try:\n",
    "        print('{:15}'.format(col), \n",
    "          'Skewness: {:05.2f}'.format(data[col].skew()) , \n",
    "          '   ' ,\n",
    "          'Kurtosis: {:06.2f}'.format(data[col].kurt())  \n",
    "         )\n",
    "    except:\n",
    "        # 无法计算出偏度和峰值的特征直接忽略\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99af7e42",
   "metadata": {},
   "source": [
    "### 步骤 10: 计算统计量，作为新特征加入数据表    \n",
    "计算某品牌的销售统计量，同学们还可以计算其他特征的统计量。    \n",
    "这里要以 train 的数据计算统计量。    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "81571297",
   "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>SaleID</th>\n",
       "      <th>name</th>\n",
       "      <th>regDate</th>\n",
       "      <th>model</th>\n",
       "      <th>brand</th>\n",
       "      <th>bodyType</th>\n",
       "      <th>fuelType</th>\n",
       "      <th>gearbox</th>\n",
       "      <th>power</th>\n",
       "      <th>kilometer</th>\n",
       "      <th>...</th>\n",
       "      <th>train</th>\n",
       "      <th>used_time</th>\n",
       "      <th>city</th>\n",
       "      <th>brand_amount</th>\n",
       "      <th>brand_price_max</th>\n",
       "      <th>brand_price_median</th>\n",
       "      <th>brand_price_min</th>\n",
       "      <th>brand_price_sum</th>\n",
       "      <th>brand_price_std</th>\n",
       "      <th>brand_price_average</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>736</td>\n",
       "      <td>20040402</td>\n",
       "      <td>30.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>4385.0</td>\n",
       "      <td>1</td>\n",
       "      <td>10217.0</td>\n",
       "      <td>59900.0</td>\n",
       "      <td>1800.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>36902172.0</td>\n",
       "      <td>4681.293524</td>\n",
       "      <td>3611.49</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2262</td>\n",
       "      <td>20030301</td>\n",
       "      <td>40.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>4757.0</td>\n",
       "      <td>4</td>\n",
       "      <td>13794.0</td>\n",
       "      <td>99900.0</td>\n",
       "      <td>6499.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>127916065.0</td>\n",
       "      <td>9369.631497</td>\n",
       "      <td>9272.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>14874</td>\n",
       "      <td>20040403</td>\n",
       "      <td>115.0</td>\n",
       "      <td>15</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>163.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>4382.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1458.0</td>\n",
       "      <td>45000.0</td>\n",
       "      <td>8500.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>14373814.0</td>\n",
       "      <td>5425.058140</td>\n",
       "      <td>9851.83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>71865</td>\n",
       "      <td>19960908</td>\n",
       "      <td>109.0</td>\n",
       "      <td>10</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>193.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>7125.0</td>\n",
       "      <td></td>\n",
       "      <td>14249.0</td>\n",
       "      <td>98000.0</td>\n",
       "      <td>5400.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>120700489.0</td>\n",
       "      <td>8988.307535</td>\n",
       "      <td>8470.21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>111080</td>\n",
       "      <td>20120103</td>\n",
       "      <td>110.0</td>\n",
       "      <td>5</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1531.0</td>\n",
       "      <td>6</td>\n",
       "      <td>4665.0</td>\n",
       "      <td>31500.0</td>\n",
       "      <td>2300.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>15424120.0</td>\n",
       "      <td>3343.624586</td>\n",
       "      <td>3305.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199995</th>\n",
       "      <td>199995</td>\n",
       "      <td>20903</td>\n",
       "      <td>19960503</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>116.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>7261.0</td>\n",
       "      <td>3</td>\n",
       "      <td>16737.0</td>\n",
       "      <td>99999.0</td>\n",
       "      <td>6000.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>142913696.0</td>\n",
       "      <td>8472.393191</td>\n",
       "      <td>8538.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199996</th>\n",
       "      <td>199996</td>\n",
       "      <td>708</td>\n",
       "      <td>19991011</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>6014.0</td>\n",
       "      <td>1</td>\n",
       "      <td>31480.0</td>\n",
       "      <td>68500.0</td>\n",
       "      <td>3199.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>174253290.0</td>\n",
       "      <td>6272.522119</td>\n",
       "      <td>5535.19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199997</th>\n",
       "      <td>199997</td>\n",
       "      <td>6693</td>\n",
       "      <td>20040412</td>\n",
       "      <td>49.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>224.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>4345.0</td>\n",
       "      <td>3</td>\n",
       "      <td>13794.0</td>\n",
       "      <td>99900.0</td>\n",
       "      <td>6499.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>127916065.0</td>\n",
       "      <td>9369.631497</td>\n",
       "      <td>9272.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199998</th>\n",
       "      <td>199998</td>\n",
       "      <td>96900</td>\n",
       "      <td>20020008</td>\n",
       "      <td>27.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>334.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>13794.0</td>\n",
       "      <td>99900.0</td>\n",
       "      <td>6499.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>127916065.0</td>\n",
       "      <td>9369.631497</td>\n",
       "      <td>9272.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199999</th>\n",
       "      <td>199999</td>\n",
       "      <td>193384</td>\n",
       "      <td>20041109</td>\n",
       "      <td>166.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>4151.0</td>\n",
       "      <td>3</td>\n",
       "      <td>10217.0</td>\n",
       "      <td>59900.0</td>\n",
       "      <td>1800.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>36902172.0</td>\n",
       "      <td>4681.293524</td>\n",
       "      <td>3611.49</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>200000 rows × 41 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        SaleID    name   regDate  model  brand  bodyType  fuelType  gearbox  \\\n",
       "0            0     736  20040402   30.0      6       1.0       0.0      0.0   \n",
       "1            1    2262  20030301   40.0      1       2.0       0.0      0.0   \n",
       "2            2   14874  20040403  115.0     15       1.0       0.0      0.0   \n",
       "3            3   71865  19960908  109.0     10       0.0       0.0      1.0   \n",
       "4            4  111080  20120103  110.0      5       1.0       0.0      0.0   \n",
       "...        ...     ...       ...    ...    ...       ...       ...      ...   \n",
       "199995  199995   20903  19960503    4.0      4       4.0       0.0      0.0   \n",
       "199996  199996     708  19991011    0.0      0       0.0       0.0      0.0   \n",
       "199997  199997    6693  20040412   49.0      1       0.0       1.0      1.0   \n",
       "199998  199998   96900  20020008   27.0      1       0.0       0.0      1.0   \n",
       "199999  199999  193384  20041109  166.0      6       1.0       NaN      1.0   \n",
       "\n",
       "        power  kilometer  ...  train  used_time  city  brand_amount  \\\n",
       "0        60.0       12.5  ...      1     4385.0     1       10217.0   \n",
       "1         0.0       15.0  ...      1     4757.0     4       13794.0   \n",
       "2       163.0       12.5  ...      1     4382.0     2        1458.0   \n",
       "3       193.0       15.0  ...      1     7125.0             14249.0   \n",
       "4        68.0        5.0  ...      1     1531.0     6        4665.0   \n",
       "...       ...        ...  ...    ...        ...   ...           ...   \n",
       "199995  116.0       15.0  ...      0     7261.0     3       16737.0   \n",
       "199996   75.0       15.0  ...      0     6014.0     1       31480.0   \n",
       "199997  224.0       15.0  ...      0     4345.0     3       13794.0   \n",
       "199998  334.0       15.0  ...      0        NaN     1       13794.0   \n",
       "199999   68.0        9.0  ...      0     4151.0     3       10217.0   \n",
       "\n",
       "        brand_price_max  brand_price_median  brand_price_min  brand_price_sum  \\\n",
       "0               59900.0              1800.0             13.0       36902172.0   \n",
       "1               99900.0              6499.0             15.0      127916065.0   \n",
       "2               45000.0              8500.0            100.0       14373814.0   \n",
       "3               98000.0              5400.0             15.0      120700489.0   \n",
       "4               31500.0              2300.0             20.0       15424120.0   \n",
       "...                 ...                 ...              ...              ...   \n",
       "199995          99999.0              6000.0             12.0      142913696.0   \n",
       "199996          68500.0              3199.0             13.0      174253290.0   \n",
       "199997          99900.0              6499.0             15.0      127916065.0   \n",
       "199998          99900.0              6499.0             15.0      127916065.0   \n",
       "199999          59900.0              1800.0             13.0       36902172.0   \n",
       "\n",
       "        brand_price_std  brand_price_average  \n",
       "0           4681.293524              3611.49  \n",
       "1           9369.631497              9272.64  \n",
       "2           5425.058140              9851.83  \n",
       "3           8988.307535              8470.21  \n",
       "4           3343.624586              3305.64  \n",
       "...                 ...                  ...  \n",
       "199995      8472.393191              8538.28  \n",
       "199996      6272.522119              5535.19  \n",
       "199997      9369.631497              9272.64  \n",
       "199998      9369.631497              9272.64  \n",
       "199999      4681.293524              3611.49  \n",
       "\n",
       "[200000 rows x 41 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_gb = train.groupby(\"brand\")\n",
    "all_info = {}\n",
    "for kind, kind_data in train_gb:\n",
    "    info = {}\n",
    "    kind_data = kind_data[kind_data['price'] > 0]\n",
    "    info['brand_amount'] = len(kind_data)\n",
    "    info['brand_price_max'] = kind_data.price.max()\n",
    "    info['brand_price_median'] = kind_data.price.median()\n",
    "    info['brand_price_min'] = kind_data.price.min()\n",
    "    info['brand_price_sum'] = kind_data.price.sum()\n",
    "    info['brand_price_std'] = kind_data.price.std()\n",
    "    info['brand_price_average'] = round(kind_data.price.sum() / (len(kind_data) + 1), 2)\n",
    "    all_info[kind] = info\n",
    "brand_fe = pd.DataFrame(all_info).T.reset_index().rename(columns={\"index\": \"brand\"})\n",
    "data = data.merge(brand_fe, how='left', on='brand')\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7e8fd08c",
   "metadata": {},
   "source": [
    "### 步骤 11: 数据离散化    \n",
    "    \n",
    "1. 离散后稀疏向量内积乘法运算速度更快，计算结果也方便存储，容易扩展；\n",
    "2. 离散后的特征对异常值更具鲁棒性，如 age>30 为 1 否则为 0，对于年龄为 200 的也不会对模型造成很大的干扰；\n",
    "3. 离散化，能够提升模型的表达能力，加大拟合；\n",
    "4. 离散后特征可以进行特征交叉，提升表达能力，由 M+N 个变量编程 M*N 个变量，进一步引入非线形，提升了表达能力；\n",
    "5. 特征离散后模型更稳定，如用户年龄区间，不会因为用户年龄长了一岁就变化。\n",
    "\n",
    "此处以 power 为例。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "a9233886",
   "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>power_bin</th>\n",
       "      <th>power</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.0</td>\n",
       "      <td>60.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>16.0</td>\n",
       "      <td>163.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>19.0</td>\n",
       "      <td>193.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6.0</td>\n",
       "      <td>68.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   power_bin  power\n",
       "0        5.0   60.0\n",
       "1        NaN    0.0\n",
       "2       16.0  163.0\n",
       "3       19.0  193.0\n",
       "4        6.0   68.0"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bin = [i*10 for i in range(31)]\n",
    "data['power_bin'] = pd.cut(data['power'], bin, labels=False)\n",
    "data[['power_bin', 'power']].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ed99d67",
   "metadata": {},
   "source": [
    "### 步骤 12: 删除无用原始特征    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "55b145ec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['SaleID', 'name', 'regDate', 'model', 'brand', 'bodyType', 'fuelType',\n",
       "       'gearbox', 'power', 'kilometer', 'notRepairedDamage', 'regionCode',\n",
       "       'seller', 'offerType', 'creatDate', 'price', 'v_0', 'v_1', 'v_2', 'v_3',\n",
       "       'v_4', 'v_5', 'v_6', 'v_7', 'v_8', 'v_9', 'v_10', 'v_11', 'v_12',\n",
       "       'v_13', 'v_14', 'train', 'used_time', 'city', 'brand_amount_x',\n",
       "       'brand_price_max_x', 'brand_price_median_x', 'brand_price_min_x',\n",
       "       'brand_price_sum_x', 'brand_price_std_x', 'brand_price_average_x',\n",
       "       'brand_amount_y', 'brand_price_max_y', 'brand_price_median_y',\n",
       "       'brand_price_min_y', 'brand_price_sum_y', 'brand_price_std_y',\n",
       "       'brand_price_average_y', 'power_bin'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 查看现有特征\n",
    "data.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "23d9715c",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_FE = data.drop(['SaleID', 'regDate', 'seller', 'offerType', \n",
    "                  'creatDate', 'regionCode'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "5aaa16f7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['name', 'model', 'brand', 'bodyType', 'fuelType', 'gearbox', 'power',\n",
       "       'kilometer', 'notRepairedDamage', 'price', 'v_0', 'v_1', 'v_2', 'v_3',\n",
       "       'v_4', 'v_5', 'v_6', 'v_7', 'v_8', 'v_9', 'v_10', 'v_11', 'v_12',\n",
       "       'v_13', 'v_14', 'train', 'used_time', 'city', 'brand_amount_x',\n",
       "       'brand_price_max_x', 'brand_price_median_x', 'brand_price_min_x',\n",
       "       'brand_price_sum_x', 'brand_price_std_x', 'brand_price_average_x',\n",
       "       'brand_amount_y', 'brand_price_max_y', 'brand_price_median_y',\n",
       "       'brand_price_min_y', 'brand_price_sum_y', 'brand_price_std_y',\n",
       "       'brand_price_average_y', 'power_bin'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 查看留存特征\n",
    "data_FE.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb93ec3a",
   "metadata": {},
   "source": [
    "### 步骤 13: 保存完成特征处理的数据  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "25f1a203",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_FE.to_csv('used_car_data_FE.csv',index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aecd1835",
   "metadata": {},
   "source": [
    "## 【任务总结】\n",
    "\n",
    "特征工程主要目的在于将数据转换为能更好地表示潜在问题的特征，从而提高机器学习的性能。比如，异常值处理是为了去除噪声，填补缺失值可以加入先验知识等。特征构造也属于特征工程的一部分，其目的是为了增强数据的表达。    \n",
    "有些特征是匿名特征，这导致我们并不清楚特征相互直接的关联性，这时我们就只有单纯基于特征进行处理，比如装箱，groupby，agg 等这样一些操作进行一些特征统计，此外还可以对特征进行进一步的 log，exp 等变换，或者对多个特征进行四则运算（如上面我们算出的使用时长），多项式组合等然后进行筛选。由于特性的匿名性其实限制了很多对于特征的处理，当然有些时候用 NN 去提取一些特征也会达到意想不到的良好效果。    \n",
    "对于知道特征含义（非匿名）的特征工程，特别是在生产实际中，会基于信号处理，频域提取，丰度，偏度等构建更为有实际意义的特征，这就是结合背景的特征构建。这样一种特征构建往往要深入分析背后的业务逻辑或者说物理原理，才能提升预测效果。    \n",
    "当然特征工程其实是和模型结合在一起的，需要对模型输入数据的类型、分布等有较多的了解。而对于特征的处理效果和特征重要性等往往要通过模型来验证。\n",
    "总的来说，特征工程是入门简单，但精通非常难的一件事。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "559b80d3",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}