Untitled.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 项目内容：\n",
    "    - 1、利用网络爬虫获取厦门市思明区二手房信息\n",
    "    - 2、进行数据预处理\n",
    "    - 3、进行数据可视化\n",
    "    - 4、构建机器学习房价预测模型\n",
    "    - 5、预测房价"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1、房屋信息的爬取"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "from bs4 import BeautifulSoup\n",
    "# requests返回网页内容\n",
    "res = requests.get(r'https://xm.esf.fang.com/house-a0352/')\n",
    "#res.text\n",
    "\n",
    "# BeautifulSoup解析网页\n",
    "soup = BeautifulSoup(res.text,'html.parser') # 使用 HTML.parser 解析器"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'户型': '3室2厅2卫', '建筑面积': '99平米', '单价': '44444元/平米', '朝向': '南北', '楼层（共7层）': '中层', '装修': '中装修', '小区名称': '前埔北区一里', '房屋总价': '440万'}\n"
     ]
    }
   ],
   "source": [
    "def get_house(url):\n",
    "    '获取页面中每个房子的信息'\n",
    "    information = {} # 存储房屋所有信息\n",
    "    res = requests.get(url)\n",
    "    soup = BeautifulSoup(res.text,'html.parser')\n",
    "    \n",
    "    # 获取户型、建筑面积、单价、朝向、楼层、装修情况\n",
    "    houses = soup.select('.tab-cont-right .trl-item1') \n",
    "    #print(houses)\n",
    "    for house in houses:\n",
    "        me = house.text.strip().split('\\n')\n",
    "        information[me[1]] = me[0].strip()\n",
    "    \n",
    "    # 获取小区名字\n",
    "    name = soup.select('.rcont .blue')\n",
    "    information['小区名称'] = name[0].text\n",
    "    \n",
    "    # 获取房屋总价\n",
    "    price = soup.select('.trl-item')\n",
    "    information['房屋总价'] = price[0].text\n",
    "    print(information)\n",
    "\n",
    "# 函数测试\n",
    "get_house(\"https://xm.esf.fang.com/chushou/3_346676334.htm?channel=1,2&psid=1_15_60\") # 每一页中房子的url(注意：链接可能失效)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "https://xm.esf.fang.com/house-a0352/i31/\n"
     ]
    }
   ],
   "source": [
    "def get_page(n):\n",
    "    '分页爬取数据'\n",
    "    for i in range(1,n+1): # n：爬取页数\n",
    "        url = r'https://xm.esf.fang.com/house-a0352/i3{}/'.format(i) # 总共多少页\n",
    "        res = requests.get(url)\n",
    "        houses = BeautifulSoup(res.text,'html.parser')\n",
    "        print(url)\n",
    "        j  = 1\n",
    "        houses = houses.select('.shop_list .clearfix h4 a') # 获取每套房子的href\n",
    "        for house in houses:\n",
    "            try:\n",
    "                demo_url = house['href']\n",
    "                url = r'https://xm.esf.fang.com' + demo_url + '?channel=1,2&psid=1_{}_60'.format(j) # 每一页中有多少套房子\n",
    "                # 获取当前页面中每套房子的信息\n",
    "                #get_house(url) \n",
    "                j += 1\n",
    "            except Exception as e:\n",
    "                print('-------->',e)\n",
    "get_page(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 房屋信息的归一化与存储"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'户型': '3室2厅2卫',\n",
       " '建筑面积': '99平米',\n",
       " '单价': '44444元/平米',\n",
       " '朝向': '南北',\n",
       " '楼层': '中层',\n",
       " '装修': '中装修',\n",
       " '小区名称': '前埔北区一里',\n",
       " '房屋总价': '440万'}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import requests\n",
    "from bs4 import BeautifulSoup\n",
    "\n",
    "def get_house(url):\n",
    "    '获取页面中每个房子的信息'\n",
    "    information = {} # 存储房屋所有信息\n",
    "    res = requests.get(url)\n",
    "    soup = BeautifulSoup(res.text,'html.parser')\n",
    "    \n",
    "    # 获取户型、建筑面积、单价、朝向、楼层、装修情况\n",
    "    houses = soup.select('.tab-cont-right .trl-item1') \n",
    "    for house in houses:\n",
    "        m = house.text.strip().split('\\n')\n",
    "        me = m[1]\n",
    "        if '朝向' in me:\n",
    "            me = me.strip('进门') \n",
    "        if '楼层'in me:\n",
    "            me = me[0:2]\n",
    "        if '地上层数' in me:\n",
    "            me = '楼层'\n",
    "        if '装修程度' in me:\n",
    "            me = '装修'\n",
    "        information[me] = m[0].strip()\n",
    "    \n",
    "    # 获取小区名字\n",
    "    name = soup.select('.rcont .blue')\n",
    "    information['小区名称'] = name[0].text\n",
    "    \n",
    "    # 获取房屋总价\n",
    "    price = soup.select('.trl-item')\n",
    "    information['房屋总价'] = price[0].text\n",
    "    return information\n",
    "\n",
    "# 函数测试\n",
    "get_house(\"https://xm.esf.fang.com/chushou/3_346676334.htm?channel=1,2&psid=1_15_60\") # 每一页中房子的url"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import time\n",
    "\n",
    "def get_page(i):\n",
    "    '分页爬取数据'\n",
    "    url = r'https://xm.esf.fang.com/house-a0352/i3{}/'.format(i) # 总共多少页\n",
    "    res = requests.get(url)\n",
    "    houses = BeautifulSoup(res.text,'html.parser')\n",
    "    #print(url)\n",
    "    j  = 1\n",
    "    houses = houses.select('.shop_list .clearfix h4 a')\n",
    "    page_information = [] # 数据存储\n",
    "    for house in houses:\n",
    "        try:\n",
    "            demo_url = house['href']\n",
    "            url = r'https://xm.esf.fang.com' + demo_url + '?channel=1,2&psid=1_{}_60'.format(j) # 每一页中有多少套房子\n",
    "            # 获取当前页中每套房子的信息\n",
    "            information = get_house(url)\n",
    "            print('正在爬取第{}页第{}套房子···'.format(i,j),end='\\r')\n",
    "            page_information.append(information)\n",
    "            j += 1\n",
    "            time.sleep(0.5) # 预防爬取频繁，防止ip被封\n",
    "        except Exception as e:\n",
    "            print('-------->',e)\n",
    "    #将爬取的数据转换为DataFrame格式\n",
    "    df = pd.DataFrame(page_information)\n",
    "    #df.to_csv('house_information.csv')\n",
    "    return df\n",
    "#get_page(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "ename": "IndentationError",
     "evalue": "expected an indented block (<ipython-input-24-21c92205b370>, line 10)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"<ipython-input-24-21c92205b370>\"\u001b[1;36m, line \u001b[1;32m10\u001b[0m\n\u001b[1;33m    except Exception as e:\u001b[0m\n\u001b[1;37m         ^\u001b[0m\n\u001b[1;31mIndentationError\u001b[0m\u001b[1;31m:\u001b[0m expected an indented block\n"
     ]
    }
   ],
   "source": [
    "# 正式爬取数据并保存为csv数据\n",
    "\n",
    "df = pd.DataFrame() # 创建一个空的DataFrame\n",
    "name_csv = 'house_information_'\n",
    "for i in range(1,101): # 总共爬取100页数据\n",
    "    try:\n",
    "        df_get = get_page(i)\n",
    "        df = df.append(df_get)\n",
    "        print(df)\n",
    "    except Exception as e:\n",
    "        print('------->',e)\n",
    "    if i/100 == 1:\n",
    "        df.to_csv(name_csv+str(i)+'.csv')\n",
    "        df = pd.DataFrame() # 清空当前爬取完成的数据，防止内存溢出"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2、数据预处理"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>index</th>\n",
       "      <th>单价</th>\n",
       "      <th>小区名称</th>\n",
       "      <th>建筑面积</th>\n",
       "      <th>户型</th>\n",
       "      <th>房屋总价</th>\n",
       "      <th>朝向</th>\n",
       "      <th>楼层</th>\n",
       "      <th>装修</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>41117元/平米</td>\n",
       "      <td>仙岳山庄</td>\n",
       "      <td>104.58平米</td>\n",
       "      <td>2室2厅2卫</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>低层</td>\n",
       "      <td>简装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>63489元/平米</td>\n",
       "      <td>禹洲华侨海景城</td>\n",
       "      <td>201.61平米</td>\n",
       "      <td>5室2厅2卫</td>\n",
       "      <td>1280万</td>\n",
       "      <td>东北</td>\n",
       "      <td>高层</td>\n",
       "      <td>豪华装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>58339元/平米</td>\n",
       "      <td>汇丰家园</td>\n",
       "      <td>128.56平米</td>\n",
       "      <td>3室2厅2卫</td>\n",
       "      <td>750万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>中装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>46739元/平米</td>\n",
       "      <td>嘉盛豪园</td>\n",
       "      <td>92平米</td>\n",
       "      <td>3室2厅1卫</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>精装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>43952元/平米</td>\n",
       "      <td>金帝花园</td>\n",
       "      <td>118.31平米</td>\n",
       "      <td>3室2厅2卫</td>\n",
       "      <td>520万</td>\n",
       "      <td>南北</td>\n",
       "      <td>高层</td>\n",
       "      <td>简装修</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   index         单价     小区名称      建筑面积      户型   房屋总价  朝向  楼层    装修\n",
       "0      0  41117元/平米     仙岳山庄  104.58平米  2室2厅2卫   430万  南北  低层   简装修\n",
       "1      1  63489元/平米  禹洲华侨海景城  201.61平米  5室2厅2卫  1280万  东北  高层  豪华装修\n",
       "2      2  58339元/平米     汇丰家园  128.56平米  3室2厅2卫   750万  南北  中层   中装修\n",
       "3      3  46739元/平米     嘉盛豪园      92平米  3室2厅1卫   430万  南北  中层   精装修\n",
       "4      4  43952元/平米     金帝花园  118.31平米  3室2厅2卫   520万  南北  高层   简装修"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv(r'house_information.csv') # 如果用pandas打不开数据，可以使用记事本打开把编码格式改成utf-8另存\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>单价</th>\n",
       "      <th>小区名称</th>\n",
       "      <th>建筑面积</th>\n",
       "      <th>户型</th>\n",
       "      <th>房屋总价</th>\n",
       "      <th>朝向</th>\n",
       "      <th>楼层</th>\n",
       "      <th>装修</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>41117元/平米</td>\n",
       "      <td>仙岳山庄</td>\n",
       "      <td>104.58平米</td>\n",
       "      <td>2室2厅2卫</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>低层</td>\n",
       "      <td>简装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>63489元/平米</td>\n",
       "      <td>禹洲华侨海景城</td>\n",
       "      <td>201.61平米</td>\n",
       "      <td>5室2厅2卫</td>\n",
       "      <td>1280万</td>\n",
       "      <td>东北</td>\n",
       "      <td>高层</td>\n",
       "      <td>豪华装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>58339元/平米</td>\n",
       "      <td>汇丰家园</td>\n",
       "      <td>128.56平米</td>\n",
       "      <td>3室2厅2卫</td>\n",
       "      <td>750万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>中装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>46739元/平米</td>\n",
       "      <td>嘉盛豪园</td>\n",
       "      <td>92平米</td>\n",
       "      <td>3室2厅1卫</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>精装修</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>43952元/平米</td>\n",
       "      <td>金帝花园</td>\n",
       "      <td>118.31平米</td>\n",
       "      <td>3室2厅2卫</td>\n",
       "      <td>520万</td>\n",
       "      <td>南北</td>\n",
       "      <td>高层</td>\n",
       "      <td>简装修</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          单价     小区名称      建筑面积      户型   房屋总价  朝向  楼层    装修\n",
       "0  41117元/平米     仙岳山庄  104.58平米  2室2厅2卫   430万  南北  低层   简装修\n",
       "1  63489元/平米  禹洲华侨海景城  201.61平米  5室2厅2卫  1280万  东北  高层  豪华装修\n",
       "2  58339元/平米     汇丰家园  128.56平米  3室2厅2卫   750万  南北  中层   中装修\n",
       "3  46739元/平米     嘉盛豪园      92平米  3室2厅1卫   430万  南北  中层   精装修\n",
       "4  43952元/平米     金帝花园  118.31平米  3室2厅2卫   520万  南北  高层   简装修"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.drop('index',axis=1,inplace=True) # 删除index列（用del更方便）\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Series的extract支持正则匹配抽取，返回的值是字符串\n",
    "data[['室','厅','卫']] = data['户型'].str.extract(r'(\\d+)室(\\d+)厅(\\d+)卫')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>单价</th>\n",
       "      <th>小区名称</th>\n",
       "      <th>建筑面积</th>\n",
       "      <th>房屋总价</th>\n",
       "      <th>朝向</th>\n",
       "      <th>楼层</th>\n",
       "      <th>装修</th>\n",
       "      <th>室</th>\n",
       "      <th>厅</th>\n",
       "      <th>卫</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>41117元/平米</td>\n",
       "      <td>仙岳山庄</td>\n",
       "      <td>104.58平米</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>低层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>63489元/平米</td>\n",
       "      <td>禹洲华侨海景城</td>\n",
       "      <td>201.61平米</td>\n",
       "      <td>1280万</td>\n",
       "      <td>东北</td>\n",
       "      <td>高层</td>\n",
       "      <td>豪华装修</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>58339元/平米</td>\n",
       "      <td>汇丰家园</td>\n",
       "      <td>128.56平米</td>\n",
       "      <td>750万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>中装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>46739元/平米</td>\n",
       "      <td>嘉盛豪园</td>\n",
       "      <td>92平米</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>精装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>43952元/平米</td>\n",
       "      <td>金帝花园</td>\n",
       "      <td>118.31平米</td>\n",
       "      <td>520万</td>\n",
       "      <td>南北</td>\n",
       "      <td>高层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          单价     小区名称      建筑面积   房屋总价  朝向  楼层    装修    室    厅    卫\n",
       "0  41117元/平米     仙岳山庄  104.58平米   430万  南北  低层   简装修  2.0  2.0  2.0\n",
       "1  63489元/平米  禹洲华侨海景城  201.61平米  1280万  东北  高层  豪华装修  5.0  2.0  2.0\n",
       "2  58339元/平米     汇丰家园  128.56平米   750万  南北  中层   中装修  3.0  2.0  2.0\n",
       "3  46739元/平米     嘉盛豪园      92平米   430万  南北  中层   精装修  3.0  2.0  1.0\n",
       "4  43952元/平米     金帝花园  118.31平米   520万  南北  高层   简装修  3.0  2.0  2.0"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 把字符串格式转化为float，并删除户型\n",
    "data['室'] = data['室'].astype(float)\n",
    "data['厅'] = data['厅'].astype(float)\n",
    "data['卫'] = data['卫'].astype(float)\n",
    "del data['户型']\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>单价</th>\n",
       "      <th>小区名称</th>\n",
       "      <th>建筑面积</th>\n",
       "      <th>房屋总价</th>\n",
       "      <th>朝向</th>\n",
       "      <th>楼层</th>\n",
       "      <th>装修</th>\n",
       "      <th>室</th>\n",
       "      <th>厅</th>\n",
       "      <th>卫</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>41117元/平米</td>\n",
       "      <td>仙岳山庄</td>\n",
       "      <td>104.58</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>低层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>63489元/平米</td>\n",
       "      <td>禹洲华侨海景城</td>\n",
       "      <td>201.61</td>\n",
       "      <td>1280万</td>\n",
       "      <td>东北</td>\n",
       "      <td>高层</td>\n",
       "      <td>豪华装修</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>58339元/平米</td>\n",
       "      <td>汇丰家园</td>\n",
       "      <td>128.56</td>\n",
       "      <td>750万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>中装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>46739元/平米</td>\n",
       "      <td>嘉盛豪园</td>\n",
       "      <td>92.00</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>精装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>43952元/平米</td>\n",
       "      <td>金帝花园</td>\n",
       "      <td>118.31</td>\n",
       "      <td>520万</td>\n",
       "      <td>南北</td>\n",
       "      <td>高层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          单价     小区名称    建筑面积   房屋总价  朝向  楼层    装修    室    厅    卫\n",
       "0  41117元/平米     仙岳山庄  104.58   430万  南北  低层   简装修  2.0  2.0  2.0\n",
       "1  63489元/平米  禹洲华侨海景城  201.61  1280万  东北  高层  豪华装修  5.0  2.0  2.0\n",
       "2  58339元/平米     汇丰家园  128.56   750万  南北  中层   中装修  3.0  2.0  2.0\n",
       "3  46739元/平米     嘉盛豪园   92.00   430万  南北  中层   精装修  3.0  2.0  1.0\n",
       "4  43952元/平米     金帝花园  118.31   520万  南北  高层   简装修  3.0  2.0  2.0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 将建筑面积后的平方米去除，并将数据类型改成浮点型\n",
    "data['建筑面积'] = data['建筑面积'].map(lambda e:e.replace('平米',''))# Series中的map\n",
    "data['建筑面积'] = data['建筑面积'].astype(float)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>单价</th>\n",
       "      <th>小区名称</th>\n",
       "      <th>建筑面积</th>\n",
       "      <th>房屋总价</th>\n",
       "      <th>朝向</th>\n",
       "      <th>楼层</th>\n",
       "      <th>装修</th>\n",
       "      <th>室</th>\n",
       "      <th>厅</th>\n",
       "      <th>卫</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>41117.0</td>\n",
       "      <td>仙岳山庄</td>\n",
       "      <td>104.58</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>低层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>63489.0</td>\n",
       "      <td>禹洲华侨海景城</td>\n",
       "      <td>201.61</td>\n",
       "      <td>1280万</td>\n",
       "      <td>东北</td>\n",
       "      <td>高层</td>\n",
       "      <td>豪华装修</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>58339.0</td>\n",
       "      <td>汇丰家园</td>\n",
       "      <td>128.56</td>\n",
       "      <td>750万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>中装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>46739.0</td>\n",
       "      <td>嘉盛豪园</td>\n",
       "      <td>92.00</td>\n",
       "      <td>430万</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>精装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>43952.0</td>\n",
       "      <td>金帝花园</td>\n",
       "      <td>118.31</td>\n",
       "      <td>520万</td>\n",
       "      <td>南北</td>\n",
       "      <td>高层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        单价     小区名称    建筑面积   房屋总价  朝向  楼层    装修    室    厅    卫\n",
       "0  41117.0     仙岳山庄  104.58   430万  南北  低层   简装修  2.0  2.0  2.0\n",
       "1  63489.0  禹洲华侨海景城  201.61  1280万  东北  高层  豪华装修  5.0  2.0  2.0\n",
       "2  58339.0     汇丰家园  128.56   750万  南北  中层   中装修  3.0  2.0  2.0\n",
       "3  46739.0     嘉盛豪园   92.00   430万  南北  中层   精装修  3.0  2.0  1.0\n",
       "4  43952.0     金帝花园  118.31   520万  南北  高层   简装修  3.0  2.0  2.0"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 将单价后的元/平米去除，并将数据类型改成浮点型\n",
    "data['单价'] = data['单价'].map(lambda e:e.replace(r'元/平米',''))\n",
    "data['单价'] = data['单价'].astype(float)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>单价</th>\n",
       "      <th>小区名称</th>\n",
       "      <th>建筑面积</th>\n",
       "      <th>房屋总价</th>\n",
       "      <th>朝向</th>\n",
       "      <th>楼层</th>\n",
       "      <th>装修</th>\n",
       "      <th>室</th>\n",
       "      <th>厅</th>\n",
       "      <th>卫</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>41117.0</td>\n",
       "      <td>仙岳山庄</td>\n",
       "      <td>104.58</td>\n",
       "      <td>430.0</td>\n",
       "      <td>南北</td>\n",
       "      <td>低层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>63489.0</td>\n",
       "      <td>禹洲华侨海景城</td>\n",
       "      <td>201.61</td>\n",
       "      <td>1280.0</td>\n",
       "      <td>东北</td>\n",
       "      <td>高层</td>\n",
       "      <td>豪华装修</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>58339.0</td>\n",
       "      <td>汇丰家园</td>\n",
       "      <td>128.56</td>\n",
       "      <td>750.0</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>中装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>46739.0</td>\n",
       "      <td>嘉盛豪园</td>\n",
       "      <td>92.00</td>\n",
       "      <td>430.0</td>\n",
       "      <td>南北</td>\n",
       "      <td>中层</td>\n",
       "      <td>精装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>43952.0</td>\n",
       "      <td>金帝花园</td>\n",
       "      <td>118.31</td>\n",
       "      <td>520.0</td>\n",
       "      <td>南北</td>\n",
       "      <td>高层</td>\n",
       "      <td>简装修</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        单价     小区名称    建筑面积    房屋总价  朝向  楼层    装修    室    厅    卫\n",
       "0  41117.0     仙岳山庄  104.58   430.0  南北  低层   简装修  2.0  2.0  2.0\n",
       "1  63489.0  禹洲华侨海景城  201.61  1280.0  东北  高层  豪华装修  5.0  2.0  2.0\n",
       "2  58339.0     汇丰家园  128.56   750.0  南北  中层   中装修  3.0  2.0  2.0\n",
       "3  46739.0     嘉盛豪园   92.00   430.0  南北  中层   精装修  3.0  2.0  1.0\n",
       "4  43952.0     金帝花园  118.31   520.0  南北  高层   简装修  3.0  2.0  2.0"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 将房屋总价后的万去除，并将数据类型改成浮点型\n",
    "data['房屋总价'] = data['房屋总价'].map(lambda e:e.replace('万',''))\n",
    "data['房屋总价'] = data['房屋总价'].astype(float)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>东</th>\n",
       "      <th>东北</th>\n",
       "      <th>东南</th>\n",
       "      <th>东西</th>\n",
       "      <th>北</th>\n",
       "      <th>南</th>\n",
       "      <th>南北</th>\n",
       "      <th>暂无</th>\n",
       "      <th>西</th>\n",
       "      <th>西北</th>\n",
       "      <th>西南</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\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>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\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>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   东  东北  东南  东西  北  南  南北  暂无  西  西北  西南\n",
       "0  0   0   0   0  0  0   1   0  0   0   0\n",
       "1  0   1   0   0  0  0   0   0  0   0   0\n",
       "2  0   0   0   0  0  0   1   0  0   0   0\n",
       "3  0   0   0   0  0  0   1   0  0   0   0\n",
       "4  0   0   0   0  0  0   1   0  0   0   0"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 使用pd.get_dummies() 量化数据\n",
    "data_direction = pd.get_dummies(data['朝向'])\n",
    "data_direction.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>中层</th>\n",
       "      <th>低层</th>\n",
       "      <th>高层</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   中层  低层  高层\n",
       "0   0   1   0\n",
       "1   0   0   1\n",
       "2   1   0   0\n",
       "3   1   0   0\n",
       "4   0   0   1"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 使用pd.get_dummies() 量化数据\n",
    "data_floor = pd.get_dummies(data['楼层'])\n",
    "data_floor.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>中装修</th>\n",
       "      <th>暂无</th>\n",
       "      <th>毛坯</th>\n",
       "      <th>简装修</th>\n",
       "      <th>精装修</th>\n",
       "      <th>豪华装修</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   中装修  暂无  毛坯  简装修  精装修  豪华装修\n",
       "0    0   0   0    1    0     0\n",
       "1    0   0   0    0    0     1\n",
       "2    1   0   0    0    0     0\n",
       "3    0   0   0    0    1     0\n",
       "4    0   0   0    1    0     0"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 使用pd.get_dummies() 量化数据\n",
    "data_decoration = pd.get_dummies(data['装修'])\n",
    "data_decoration.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 使用pd.concat矩阵拼接，axis=1：水平拼接\n",
    "data = pd.concat([data,data_direction,data_floor,data_decoration],axis=1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['单价', '小区名称', '建筑面积', '房屋总价', '朝向', '楼层', '装修', '室', '厅', '卫', '东',\n",
       "       '东北', '东南', '东西', '北', '南', '南北', '暂无', '西', '西北', '西南', '中层', '低层',\n",
       "       '高层', '中装修', '暂无', '毛坯', '简装修', '精装修', '豪华装修'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 拼接后的列名\n",
    "data.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['单价', '建筑面积', '房屋总价', '室', '厅', '卫', '东', '东北', '东南', '北', '南', '西',\n",
       "       '西北', '西南', '低层', '高层', '毛坯', '简装修', '精装修', '豪华装修'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 特征帅选\n",
    "del data['小区名称']\n",
    "del data['朝向']\n",
    "del data['楼层']\n",
    "del data['装修']\n",
    "del data['东西']\n",
    "del data['南北']\n",
    "del data['暂无'] # 两列都删除\n",
    "del data['中层'] # 多重共线性问题（线性回归）\n",
    "del data['中装修']\n",
    "data.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>单价</th>\n",
       "      <th>建筑面积</th>\n",
       "      <th>房屋总价</th>\n",
       "      <th>室</th>\n",
       "      <th>厅</th>\n",
       "      <th>卫</th>\n",
       "      <th>东</th>\n",
       "      <th>东北</th>\n",
       "      <th>东南</th>\n",
       "      <th>北</th>\n",
       "      <th>南</th>\n",
       "      <th>西</th>\n",
       "      <th>西北</th>\n",
       "      <th>西南</th>\n",
       "      <th>低层</th>\n",
       "      <th>高层</th>\n",
       "      <th>毛坯</th>\n",
       "      <th>简装修</th>\n",
       "      <th>精装修</th>\n",
       "      <th>豪华装修</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>41117.0</td>\n",
       "      <td>104.58</td>\n",
       "      <td>430.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>63489.0</td>\n",
       "      <td>201.61</td>\n",
       "      <td>1280.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>58339.0</td>\n",
       "      <td>128.56</td>\n",
       "      <td>750.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</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>0</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>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>46739.0</td>\n",
       "      <td>92.00</td>\n",
       "      <td>430.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>43952.0</td>\n",
       "      <td>118.31</td>\n",
       "      <td>520.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        单价    建筑面积    房屋总价    室    厅    卫  东  东北  东南  北  南  西  西北  西南  低层  高层  \\\n",
       "0  41117.0  104.58   430.0  2.0  2.0  2.0  0   0   0  0  0  0   0   0   1   0   \n",
       "1  63489.0  201.61  1280.0  5.0  2.0  2.0  0   1   0  0  0  0   0   0   0   1   \n",
       "2  58339.0  128.56   750.0  3.0  2.0  2.0  0   0   0  0  0  0   0   0   0   0   \n",
       "3  46739.0   92.00   430.0  3.0  2.0  1.0  0   0   0  0  0  0   0   0   0   0   \n",
       "4  43952.0  118.31   520.0  3.0  2.0  2.0  0   0   0  0  0  0   0   0   0   1   \n",
       "\n",
       "   毛坯  简装修  精装修  豪华装修  \n",
       "0   0    1    0     0  \n",
       "1   0    0    0     1  \n",
       "2   0    0    0     0  \n",
       "3   0    0    1     0  \n",
       "4   0    1    0     0  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 5830 entries, 0 to 5829\n",
      "Data columns (total 20 columns):\n",
      "单价      5830 non-null float64\n",
      "建筑面积    5830 non-null float64\n",
      "房屋总价    5830 non-null float64\n",
      "室       5748 non-null float64\n",
      "厅       5748 non-null float64\n",
      "卫       5748 non-null float64\n",
      "东       5830 non-null uint8\n",
      "东北      5830 non-null uint8\n",
      "东南      5830 non-null uint8\n",
      "北       5830 non-null uint8\n",
      "南       5830 non-null uint8\n",
      "西       5830 non-null uint8\n",
      "西北      5830 non-null uint8\n",
      "西南      5830 non-null uint8\n",
      "低层      5830 non-null uint8\n",
      "高层      5830 non-null uint8\n",
      "毛坯      5830 non-null uint8\n",
      "简装修     5830 non-null uint8\n",
      "精装修     5830 non-null uint8\n",
      "豪华装修    5830 non-null uint8\n",
      "dtypes: float64(6), uint8(14)\n",
      "memory usage: 353.1 KB\n"
     ]
    }
   ],
   "source": [
    "data.info() # 发现 室厅卫中 有缺失值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 5748 entries, 0 to 5829\n",
      "Data columns (total 20 columns):\n",
      "单价      5748 non-null float64\n",
      "建筑面积    5748 non-null float64\n",
      "房屋总价    5748 non-null float64\n",
      "室       5748 non-null float64\n",
      "厅       5748 non-null float64\n",
      "卫       5748 non-null float64\n",
      "东       5748 non-null uint8\n",
      "东北      5748 non-null uint8\n",
      "东南      5748 non-null uint8\n",
      "北       5748 non-null uint8\n",
      "南       5748 non-null uint8\n",
      "西       5748 non-null uint8\n",
      "西北      5748 non-null uint8\n",
      "西南      5748 non-null uint8\n",
      "低层      5748 non-null uint8\n",
      "高层      5748 non-null uint8\n",
      "毛坯      5748 non-null uint8\n",
      "简装修     5748 non-null uint8\n",
      "精装修     5748 non-null uint8\n",
      "豪华装修    5748 non-null uint8\n",
      "dtypes: float64(6), uint8(14)\n",
      "memory usage: 392.9 KB\n"
     ]
    }
   ],
   "source": [
    "# 删除缺失值\n",
    "data.dropna(inplace=True)\n",
    "data.info()                       # 到这里数据预处理完毕"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3、数据可视化"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "area = data['建筑面积']\n",
    "price = data['房屋总价']\n",
    "plt.scatter(area,price)\n",
    "plt.show() # 有离群点数据，对线性分析不利，需要过滤"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = data[data['建筑面积'] <=300] # 正常住宅面积小于等于300平米\n",
    "area = df['建筑面积']\n",
    "price = df['房屋总价']\n",
    "#print(area.count()) #过滤后的数据量\n",
    "plt.scatter(area,price)\n",
    "plt.xlabel(\"area\")\n",
    "plt.ylabel(\"price\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4、构建房价预测模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[-187.73888919] [[7.39200324]]\n"
     ]
    }
   ],
   "source": [
    "# 先根据建筑面积和房屋总价训练模型（一元线性回归）\n",
    "from sklearn.linear_model import LinearRegression\n",
    "linear = LinearRegression()\n",
    "area = np.array(area).reshape(-1,1) # 这里需要注意新版的sklearn需要将数据转换为矩阵才能进行计算\n",
    "price = np.array(price).reshape(-1,1)\n",
    "# 训练模型\n",
    "model = linear.fit(area,price)\n",
    "# 打印截距和回归系数\n",
    "print(model.intercept_, model.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 线性回归可视化(数据拟合)\n",
    "linear_p = model.predict(area)\n",
    "plt.figure(figsize=(12,6))\n",
    "plt.scatter(area,price)\n",
    "plt.plot(area,linear_p,'red')\n",
    "plt.xlabel(\"area\")\n",
    "plt.ylabel(\"price\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 多特征模型训练（多元线性回归）"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols = ['建筑面积','室', '厅', '卫', '东', '东北', '东南', '北', '南', '西',\n",
    "       '西北', '西南', '低层', '高层', '毛坯', '简装修', '精装修', '豪华装修']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "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>建筑面积</th>\n",
       "      <th>室</th>\n",
       "      <th>厅</th>\n",
       "      <th>卫</th>\n",
       "      <th>东</th>\n",
       "      <th>东北</th>\n",
       "      <th>东南</th>\n",
       "      <th>北</th>\n",
       "      <th>南</th>\n",
       "      <th>西</th>\n",
       "      <th>西北</th>\n",
       "      <th>西南</th>\n",
       "      <th>低层</th>\n",
       "      <th>高层</th>\n",
       "      <th>毛坯</th>\n",
       "      <th>简装修</th>\n",
       "      <th>精装修</th>\n",
       "      <th>豪华装修</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>104.58</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>201.61</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>128.56</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</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>0</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>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>92.00</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>118.31</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     建筑面积    室    厅    卫  东  东北  东南  北  南  西  西北  西南  低层  高层  毛坯  简装修  精装修  \\\n",
       "0  104.58  2.0  2.0  2.0  0   0   0  0  0  0   0   0   1   0   0    1    0   \n",
       "1  201.61  5.0  2.0  2.0  0   1   0  0  0  0   0   0   0   1   0    0    0   \n",
       "2  128.56  3.0  2.0  2.0  0   0   0  0  0  0   0   0   0   0   0    0    0   \n",
       "3   92.00  3.0  2.0  1.0  0   0   0  0  0  0   0   0   0   0   0    0    1   \n",
       "4  118.31  3.0  2.0  2.0  0   0   0  0  0  0   0   0   0   1   0    1    0   \n",
       "\n",
       "   豪华装修  \n",
       "0     0  \n",
       "1     1  \n",
       "2     0  \n",
       "3     0  \n",
       "4     0  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df[cols]\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     430.0\n",
       "1    1280.0\n",
       "2     750.0\n",
       "3     430.0\n",
       "4     520.0\n",
       "Name: 房屋总价, dtype: float64"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = df['房屋总价']\n",
    "y.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "<class 'pandas.core.series.Series'>\n",
      "(4406, 18) (4406,)\n",
      "(1102, 18) (1102,)\n"
     ]
    }
   ],
   "source": [
    "print(type(X))\n",
    "print(type(y))\n",
    "# 使用train_test_split进行交叉验证\n",
    "from sklearn.model_selection import train_test_split\n",
    "x_train, x_test, y_train, y_test = train_test_split(X,y,test_size=0.2,random_state=12)\n",
    "print(x_train.shape,y_train.shape)\n",
    "print(x_test.shape,y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-194.1005982863286 [  6.81080385  32.66440227 -36.27370796   8.17388714  22.82870564\n",
      " -34.63169345   0.9396494   31.52349203  73.23243807  36.57946634\n",
      " -20.85576479  35.85073408   4.13506426 -10.09803905  70.75146155\n",
      "  -3.6000263   48.66249505 124.16793037]\n"
     ]
    }
   ],
   "source": [
    "# 模型训练\n",
    "linear = LinearRegression()\n",
    "model = linear.fit(x_train,y_train)\n",
    "print(model.intercept_, model.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "模型得分： 0.7639142687890216\n"
     ]
    }
   ],
   "source": [
    "# 模型性能评分\n",
    "price_end = model.predict(x_test)\n",
    "score = model.score(x_test,y_test) \n",
    "print(\"模型得分：\",score)# 一般模型在0.6以上就表现的不错"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 使用多种特征组合都可以预测房价，那么怎么找出最佳组合，这里使用假设验证法，选出最佳特征组合"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "        <td>Model:</td>               <td>OLS</td>         <td>Adj. R-squared:</td>      <td>0.771</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <td>Dependent Variable:</td>       <td>房屋总价</td>              <td>AIC:</td>         <td>74021.4015</td>\n",
       "</tr>\n",
       "<tr>\n",
       "         <td>Date:</td>        <td>2019-02-26 16:17</td>        <td>BIC:</td>         <td>74147.0667</td>\n",
       "</tr>\n",
       "<tr>\n",
       "   <td>No. Observations:</td>        <td>5508</td>         <td>Log-Likelihood:</td>     <td>-36992.</td> \n",
       "</tr>\n",
       "<tr>\n",
       "       <td>Df Model:</td>             <td>18</td>           <td>F-statistic:</td>        <td>1030.</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "     <td>Df Residuals:</td>          <td>5489</td>       <td>Prob (F-statistic):</td>    <td>0.00</td>   \n",
       "</tr>\n",
       "<tr>\n",
       "      <td>R-squared:</td>            <td>0.772</td>            <td>Scale:</td>          <td>40037.</td>  \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "    <td></td>      <th>Coef.</th>   <th>Std.Err.</th>     <th>t</th>     <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th> <td>-190.5740</td>  <td>14.3527</td> <td>-13.2779</td> <td>0.0000</td> <td>-218.7111</td> <td>-162.4370</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>建筑面积</th>   <td>6.7798</td>    <td>0.1065</td>   <td>63.6645</td> <td>0.0000</td>  <td>6.5711</td>    <td>6.9886</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>室</th>      <td>32.4766</td>   <td>4.8180</td>   <td>6.7407</td>  <td>0.0000</td>  <td>23.0315</td>   <td>41.9218</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>厅</th>     <td>-36.6403</td>   <td>7.4305</td>   <td>-4.9311</td> <td>0.0000</td> <td>-51.2070</td>  <td>-22.0735</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>卫</th>      <td>5.9103</td>    <td>5.9003</td>   <td>1.0017</td>  <td>0.3165</td>  <td>-5.6565</td>   <td>17.4771</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>东</th>      <td>23.4166</td>   <td>13.2163</td>  <td>1.7718</td>  <td>0.0765</td>  <td>-2.4925</td>   <td>49.3258</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>东北</th>    <td>-34.1670</td>   <td>17.2426</td>  <td>-1.9815</td> <td>0.0476</td> <td>-67.9694</td>   <td>-0.3645</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>东南</th>     <td>9.0724</td>    <td>10.6111</td>  <td>0.8550</td>  <td>0.3926</td> <td>-11.7295</td>   <td>29.8744</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>北</th>      <td>25.5771</td>   <td>19.4207</td>  <td>1.3170</td>  <td>0.1879</td> <td>-12.4952</td>   <td>63.6493</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>南</th>      <td>69.3747</td>   <td>11.3337</td>  <td>6.1211</td>  <td>0.0000</td>  <td>47.1560</td>   <td>91.5933</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>西</th>      <td>28.2927</td>   <td>24.8301</td>  <td>1.1395</td>  <td>0.2546</td> <td>-20.3842</td>   <td>76.9695</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>西北</th>    <td>-21.8337</td>   <td>15.8412</td>  <td>-1.3783</td> <td>0.1682</td> <td>-52.8888</td>   <td>9.2213</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>西南</th>     <td>33.8956</td>   <td>13.4285</td>  <td>2.5241</td>  <td>0.0116</td>  <td>7.5703</td>    <td>60.2208</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>低层</th>     <td>8.6321</td>    <td>6.8894</td>   <td>1.2530</td>  <td>0.2103</td>  <td>-4.8738</td>   <td>22.1381</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>高层</th>     <td>-8.0947</td>   <td>6.3467</td>   <td>-1.2754</td> <td>0.2022</td> <td>-20.5367</td>   <td>4.3472</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>毛坯</th>     <td>86.0976</td>   <td>15.2708</td>  <td>5.6381</td>  <td>0.0000</td>  <td>56.1608</td>  <td>116.0343</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>简装修</th>    <td>-1.7023</td>   <td>8.4182</td>   <td>-0.2022</td> <td>0.8398</td> <td>-18.2053</td>   <td>14.8007</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>精装修</th>    <td>51.9228</td>   <td>8.3723</td>   <td>6.2018</td>  <td>0.0000</td>  <td>35.5098</td>   <td>68.3358</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>豪华装修</th>  <td>113.2164</td>   <td>12.8088</td>  <td>8.8389</td>  <td>0.0000</td>  <td>88.1060</td>  <td>138.3267</td> \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "     <td>Omnibus:</td>    <td>1795.844</td>  <td>Durbin-Watson:</td>     <td>2.046</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <td>Prob(Omnibus):</td>   <td>0.000</td>  <td>Jarque-Bera (JB):</td> <td>16937.679</td>\n",
       "</tr>\n",
       "<tr>\n",
       "       <td>Skew:</td>       <td>1.283</td>      <td>Prob(JB):</td>       <td>0.000</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "     <td>Kurtosis:</td>    <td>11.199</td>   <td>Condition No.:</td>     <td>1177</td>   \n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary2.Summary'>\n",
       "\"\"\"\n",
       "                  Results: Ordinary least squares\n",
       "===================================================================\n",
       "Model:              OLS              Adj. R-squared:     0.771     \n",
       "Dependent Variable: 房屋总价             AIC:                74021.4015\n",
       "Date:               2019-02-26 16:17 BIC:                74147.0667\n",
       "No. Observations:   5508             Log-Likelihood:     -36992.   \n",
       "Df Model:           18               F-statistic:        1030.     \n",
       "Df Residuals:       5489             Prob (F-statistic): 0.00      \n",
       "R-squared:          0.772            Scale:              40037.    \n",
       "--------------------------------------------------------------------\n",
       "           Coef.    Std.Err.     t      P>|t|     [0.025     0.975] \n",
       "--------------------------------------------------------------------\n",
       "const    -190.5740   14.3527  -13.2779  0.0000  -218.7111  -162.4370\n",
       "建筑面积        6.7798    0.1065   63.6645  0.0000     6.5711     6.9886\n",
       "室          32.4766    4.8180    6.7407  0.0000    23.0315    41.9218\n",
       "厅         -36.6403    7.4305   -4.9311  0.0000   -51.2070   -22.0735\n",
       "卫           5.9103    5.9003    1.0017  0.3165    -5.6565    17.4771\n",
       "东          23.4166   13.2163    1.7718  0.0765    -2.4925    49.3258\n",
       "东北        -34.1670   17.2426   -1.9815  0.0476   -67.9694    -0.3645\n",
       "东南          9.0724   10.6111    0.8550  0.3926   -11.7295    29.8744\n",
       "北          25.5771   19.4207    1.3170  0.1879   -12.4952    63.6493\n",
       "南          69.3747   11.3337    6.1211  0.0000    47.1560    91.5933\n",
       "西          28.2927   24.8301    1.1395  0.2546   -20.3842    76.9695\n",
       "西北        -21.8337   15.8412   -1.3783  0.1682   -52.8888     9.2213\n",
       "西南         33.8956   13.4285    2.5241  0.0116     7.5703    60.2208\n",
       "低层          8.6321    6.8894    1.2530  0.2103    -4.8738    22.1381\n",
       "高层         -8.0947    6.3467   -1.2754  0.2022   -20.5367     4.3472\n",
       "毛坯         86.0976   15.2708    5.6381  0.0000    56.1608   116.0343\n",
       "简装修        -1.7023    8.4182   -0.2022  0.8398   -18.2053    14.8007\n",
       "精装修        51.9228    8.3723    6.2018  0.0000    35.5098    68.3358\n",
       "豪华装修      113.2164   12.8088    8.8389  0.0000    88.1060   138.3267\n",
       "-------------------------------------------------------------------\n",
       "Omnibus:             1795.844      Durbin-Watson:         2.046    \n",
       "Prob(Omnibus):       0.000         Jarque-Bera (JB):      16937.679\n",
       "Skew:                1.283         Prob(JB):              0.000    \n",
       "Kurtosis:            11.199        Condition No.:         1177     \n",
       "===================================================================\n",
       "* The condition number is large (1e+03). This might indicate\n",
       "strong multicollinearity or other numerical problems.\n",
       "\"\"\""
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 使用假设验证法，选出最佳特征组合\n",
    "cols = ['建筑面积','室', '厅', '卫', '东', '东北', '东南', '北', '南', '西',\n",
    "       '西北', '西南', '低层', '高层', '毛坯', '简装修', '精装修', '豪华装修']\n",
    "import statsmodels.api as sm\n",
    "Y = df['房屋总价']\n",
    "X = df[cols]\n",
    "X_ = sm.add_constant(X) #增加一列值为1的const列，保证偏置项的正常\n",
    "#print(X_)\n",
    "# 使用最小平方法\n",
    "result = sm.OLS(Y,X_)\n",
    "# 使用fit方法进行计算\n",
    "summary = result.fit()\n",
    "# 调用summary2方法打印出假设验证信息（性能指标）\n",
    "summary.summary2() # R-squared:模型评分 AIC：组合完越小越好"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "名词解释\n",
    "\n",
    "- coef 回归系数\n",
    "- Std.Err 标准差\n",
    "- t 虚无假设成立时的t值\n",
    "- P>|t| 虚无假设成立时的概率值\n",
    "- [0.025,0.975] 97.5%置信估计区间\n",
    "- 要做假设性验证，首先要设置显著性标准。\n",
    "- a.假设显著性标准是0.01\n",
    "- b.推翻虚无假设的标准是p<0.01\n",
    "- c.上面的SqFt的t=9.2416，P(>5) = 0.0000 < 0.01,因此虚无假设被推翻（这里的虚无假设是SqFt对price的回归系数为0，即SqFt与price不相关）\n",
    "\n",
    "F统计\n",
    "- 回归平方和Regression Square Sum[RSS]:依变量的变化归咎于回归模型A=sum((y-y_)^2\n",
    "- 误差平方和Error Square Sum[ESS]:依变量的变化归咎于线性模型B=sum((y-y_)^2\n",
    "- 总的平方和Total Square Sum[TSS]:依变量整体变化C=A+B\n",
    "- 回归平方平均Model Mean Square:=RSS/Regression d.f(k) k=自变量的数量\n",
    "- 误差平方平均Error Mean square:=ESS/Error d.f(n-k-1) n=观测值得数量\n",
    "- F统计F=Model Mean Square /Error Mean Square\n",
    "- F值越大越好，Prob(F-statistic)越小越好\n",
    "\n",
    "R Square\n",
    "- 回归可以解释变量比例，可以作为自变量预测因变量准确度的指标\n",
    "- SSE(残差平方和) = sum((y-y_)^2)\n",
    "- SST(整体平方和）= sum((yi-yavg)^2\n",
    "- R^2 = 1-SEE/SST 一般要大于0.6,0.7才算好\n",
    "\n",
    "Adjust R Square\n",
    "- R^2 = 1-SSE/SST SSE最小，推导出R^2不会递减\n",
    "- yi = b1x1 + b2x2 +...+bkxk+...增加任何一个变量都会增加R^2\n",
    "- Adj R^2 = 1-(1-R^2) * ((n-1)/(n-p-1))\n",
    "- n为总体大小，p为回归因子个数\n",
    "\n",
    "AIC/BIC\n",
    "- AIC(The Akaike Information Criterion)= 2K + nln(SSE/n) K是参数数量，n是观察数，SSE是残差平方和。\n",
    "- AIC鼓励数据拟合的优良性，但是应该尽量避免过拟合，所以优先考虑的模型应该是AIC最小的哪一个，\n",
    "- 赤池信息量的准则是寻找可以最好的解释数据但是包含最少自由参数的模型。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "import itertools\n",
    "\n",
    "list1 = [1, 2,3, 4, 5,6,7,8,9,10,11,12,13,14,15,16] #特征超过16个将发生异常\n",
    "list2 = []\n",
    "for i in range(1, len(list1)+1):\n",
    "    iter1 = itertools.combinations(list1, i)\n",
    "    list2.append(list(iter1))\n",
    "#print(list2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "import itertools\n",
    "# 使用itertools，找出AIC最小值的特征组合作为模型训练的特征\n",
    "# 寻找最小AIC值的特征组合\n",
    "fileds = ['建筑面积','室', '厅', '卫', '东','北', '南', '西','低层', '高层', '毛坯', '简装修', '精装修', '豪华装修']\n",
    "acis = {}\n",
    "for i in range(1,len(fileds)+1):\n",
    "    for virables in itertools.combinations(fileds,i): #从fileds中随机选择i个特征机型组合，返回的virables为元组类型\n",
    "        x1 = sm.add_constant(df[list(virables)])\n",
    "        x2 = sm.OLS(Y,x1)\n",
    "        res = x2.fit()\n",
    "        acis[virables] = res.aic # AIC评分越小越好"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(('建筑面积', '室', '厅', '东', '南', '西', '低层', '毛坯', '精装修', '豪华装修'),\n",
       "  74023.56464247877),\n",
       " (('建筑面积', '室', '厅', '东', '北', '南', '西', '高层', '毛坯', '精装修', '豪华装修'),\n",
       "  74023.41564227329),\n",
       " (('建筑面积', '室', '厅', '卫', '东', '南', '高层', '毛坯', '精装修', '豪华装修'),\n",
       "  74023.36778965591),\n",
       " (('建筑面积', '室', '厅', '东', '北', '南', '低层', '高层', '毛坯', '精装修', '豪华装修'),\n",
       "  74023.31766334853),\n",
       " (('建筑面积', '室', '厅', '东', '南', '西', '高层', '毛坯', '精装修', '豪华装修'),\n",
       "  74023.14955002116),\n",
       " (('建筑面积', '室', '厅', '东', '北', '南', '低层', '毛坯', '精装修', '豪华装修'),\n",
       "  74023.14910923649),\n",
       " (('建筑面积', '室', '厅', '东', '南', '低层', '高层', '毛坯', '精装修', '豪华装修'),\n",
       "  74023.02854387014),\n",
       " (('建筑面积', '室', '厅', '东', '北', '南', '高层', '毛坯', '精装修', '豪华装修'),\n",
       "  74022.91373410511),\n",
       " (('建筑面积', '室', '厅', '东', '南', '低层', '毛坯', '精装修', '豪华装修'), 74022.8722434788),\n",
       " (('建筑面积', '室', '厅', '东', '南', '高层', '毛坯', '精装修', '豪华装修'), 74022.56669895534)]"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from collections import Counter\n",
    "# 对字典进行统计\n",
    "counter = Counter(acis)\n",
    "# 降序选出AIC最小的10个数，也就是最佳特征组合\n",
    "counter.most_common()[-10:] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.764510909319327"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 接下来使用AIC值最小的特征组合进行预测\n",
    "col2 = ['建筑面积', '室', '厅', '东', '南', '高层', '毛坯', '精装修', '豪华装修']\n",
    "X = df[col2]\n",
    "y = df['房屋总价']\n",
    "x_train, x_test, y_train, y_test = train_test_split(X,y,test_size=0.2,random_state=13)\n",
    "linear = LinearRegression()\n",
    "model = linear.fit(x_train,y_train)\n",
    "model.score(x_test,y_test) # 模型性能有所提高，但是提升的不明显"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5、房价的预测"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "现在我们可以根据给定的最佳特征组合进行预测房价"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([811.49301398])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 假设我要买一套房子（想想就觉得很美），房子面积120平米，3室，1厅，南面，高层，精装修\n",
    "my_house = [120,3,1,0,1,1,0,1,0] #根据col2特征\n",
    "my_house = np.array(my_house).reshape(-1,1).T\n",
    "#print(x_test)\n",
    "model.predict(my_house)# 预测价格"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "End···"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}