LR_RentalListingInquiries.ipynb · Patrick Qi/w2 - 亓呈文 - 854510525

加入 Gitee

与超过 1200万开发者一起发现、参与优秀开源项目，私有仓库也完全免费：）

该仓库未声明开源许可证文件（LICENSE），使用请关注具体项目描述及其代码上游依赖。

克隆/下载

LR_RentalListingInquiries.ipynb 117.54 KB

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Logistic 回归 -- 租房意向预测"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 导入必要模块\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# 竞赛的评价指标为 logloss\n",
    "from sklearn.metrics import log_loss\n",
    "\n",
    "from matplotlib import pyplot\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "pd.set_option('display.float_format', lambda x: '%.3f' % x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 读入数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('RentListingInquries_FE_train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test = pd.read_csv('RentListingInquries_FE_test.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 数据探索"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>price</th>\n",
       "      <th>price_bathrooms</th>\n",
       "      <th>price_bedrooms</th>\n",
       "      <th>room_diff</th>\n",
       "      <th>room_num</th>\n",
       "      <th>Year</th>\n",
       "      <th>...</th>\n",
       "      <th>walk</th>\n",
       "      <th>walls</th>\n",
       "      <th>war</th>\n",
       "      <th>washer</th>\n",
       "      <th>water</th>\n",
       "      <th>wheelchair</th>\n",
       "      <th>wifi</th>\n",
       "      <th>windows</th>\n",
       "      <th>work</th>\n",
       "      <th>interest_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.500</td>\n",
       "      <td>3</td>\n",
       "      <td>40.715</td>\n",
       "      <td>-73.942</td>\n",
       "      <td>3000</td>\n",
       "      <td>1200.000</td>\n",
       "      <td>750.000</td>\n",
       "      <td>-1.500</td>\n",
       "      <td>4.500</td>\n",
       "      <td>2016</td>\n",
       "      <td>...</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.000</td>\n",
       "      <td>2</td>\n",
       "      <td>40.795</td>\n",
       "      <td>-73.967</td>\n",
       "      <td>5465</td>\n",
       "      <td>2732.500</td>\n",
       "      <td>1821.667</td>\n",
       "      <td>-1.000</td>\n",
       "      <td>3.000</td>\n",
       "      <td>2016</td>\n",
       "      <td>...</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>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.000</td>\n",
       "      <td>1</td>\n",
       "      <td>40.739</td>\n",
       "      <td>-74.002</td>\n",
       "      <td>2850</td>\n",
       "      <td>1425.000</td>\n",
       "      <td>1425.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>2016</td>\n",
       "      <td>...</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>1.000</td>\n",
       "      <td>1</td>\n",
       "      <td>40.754</td>\n",
       "      <td>-73.968</td>\n",
       "      <td>3275</td>\n",
       "      <td>1637.500</td>\n",
       "      <td>1637.500</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>2016</td>\n",
       "      <td>...</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>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.000</td>\n",
       "      <td>4</td>\n",
       "      <td>40.824</td>\n",
       "      <td>-73.949</td>\n",
       "      <td>3350</td>\n",
       "      <td>1675.000</td>\n",
       "      <td>670.000</td>\n",
       "      <td>-3.000</td>\n",
       "      <td>5.000</td>\n",
       "      <td>2016</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 225 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   bathrooms  bedrooms  latitude  longitude  price  price_bathrooms  \\\n",
       "0      1.500         3    40.715    -73.942   3000         1200.000   \n",
       "1      1.000         2    40.795    -73.967   5465         2732.500   \n",
       "2      1.000         1    40.739    -74.002   2850         1425.000   \n",
       "3      1.000         1    40.754    -73.968   3275         1637.500   \n",
       "4      1.000         4    40.824    -73.949   3350         1675.000   \n",
       "\n",
       "   price_bedrooms  room_diff  room_num  Year       ...        walk  walls  \\\n",
       "0         750.000     -1.500     4.500  2016       ...           0      0   \n",
       "1        1821.667     -1.000     3.000  2016       ...           0      0   \n",
       "2        1425.000      0.000     2.000  2016       ...           0      0   \n",
       "3        1637.500      0.000     2.000  2016       ...           0      0   \n",
       "4         670.000     -3.000     5.000  2016       ...           0      0   \n",
       "\n",
       "   war  washer  water  wheelchair  wifi  windows  work  interest_level  \n",
       "0    0       0      0           0     0        0     0               1  \n",
       "1    0       0      0           0     0        0     0               2  \n",
       "2    0       0      0           0     0        0     0               0  \n",
       "3    0       0      0           0     0        0     0               2  \n",
       "4    1       0      0           0     0        0     0               2  \n",
       "\n",
       "[5 rows x 225 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 49352 entries, 0 to 49351\n",
      "Columns: 225 entries, bathrooms to interest_level\n",
      "dtypes: float64(7), int64(218)\n",
      "memory usage: 84.7 MB\n"
     ]
    }
   ],
   "source": [
    "train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>price</th>\n",
       "      <th>price_bathrooms</th>\n",
       "      <th>price_bedrooms</th>\n",
       "      <th>room_diff</th>\n",
       "      <th>room_num</th>\n",
       "      <th>Year</th>\n",
       "      <th>...</th>\n",
       "      <th>walk</th>\n",
       "      <th>walls</th>\n",
       "      <th>war</th>\n",
       "      <th>washer</th>\n",
       "      <th>water</th>\n",
       "      <th>wheelchair</th>\n",
       "      <th>wifi</th>\n",
       "      <th>windows</th>\n",
       "      <th>work</th>\n",
       "      <th>interest_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>...</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "      <td>49352.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.212</td>\n",
       "      <td>1.542</td>\n",
       "      <td>40.742</td>\n",
       "      <td>-73.956</td>\n",
       "      <td>3830.174</td>\n",
       "      <td>1697.863</td>\n",
       "      <td>1657.567</td>\n",
       "      <td>-0.329</td>\n",
       "      <td>2.754</td>\n",
       "      <td>2016.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.003</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.186</td>\n",
       "      <td>0.009</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.028</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.001</td>\n",
       "      <td>1.617</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.501</td>\n",
       "      <td>1.115</td>\n",
       "      <td>0.639</td>\n",
       "      <td>1.178</td>\n",
       "      <td>22066.866</td>\n",
       "      <td>11004.769</td>\n",
       "      <td>7817.996</td>\n",
       "      <td>0.948</td>\n",
       "      <td>1.446</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.055</td>\n",
       "      <td>0.020</td>\n",
       "      <td>0.389</td>\n",
       "      <td>0.102</td>\n",
       "      <td>0.021</td>\n",
       "      <td>0.165</td>\n",
       "      <td>0.045</td>\n",
       "      <td>0.032</td>\n",
       "      <td>0.031</td>\n",
       "      <td>0.626</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>-118.271</td>\n",
       "      <td>43.000</td>\n",
       "      <td>21.500</td>\n",
       "      <td>43.000</td>\n",
       "      <td>-5.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2016.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>40.728</td>\n",
       "      <td>-73.992</td>\n",
       "      <td>2500.000</td>\n",
       "      <td>1225.000</td>\n",
       "      <td>1066.667</td>\n",
       "      <td>-1.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>2016.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>40.752</td>\n",
       "      <td>-73.978</td>\n",
       "      <td>3150.000</td>\n",
       "      <td>1500.000</td>\n",
       "      <td>1383.417</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>2016.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>40.774</td>\n",
       "      <td>-73.955</td>\n",
       "      <td>4100.000</td>\n",
       "      <td>1850.000</td>\n",
       "      <td>1962.500</td>\n",
       "      <td>0.000</td>\n",
       "      <td>4.000</td>\n",
       "      <td>2016.000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>10.000</td>\n",
       "      <td>8.000</td>\n",
       "      <td>44.883</td>\n",
       "      <td>0.000</td>\n",
       "      <td>4490000.000</td>\n",
       "      <td>2245000.000</td>\n",
       "      <td>1496666.667</td>\n",
       "      <td>8.000</td>\n",
       "      <td>13.500</td>\n",
       "      <td>2016.000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>2.000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 225 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       bathrooms  bedrooms  latitude  longitude       price  price_bathrooms  \\\n",
       "count  49352.000 49352.000 49352.000  49352.000   49352.000        49352.000   \n",
       "mean       1.212     1.542    40.742    -73.956    3830.174         1697.863   \n",
       "std        0.501     1.115     0.639      1.178   22066.866        11004.769   \n",
       "min        0.000     0.000     0.000   -118.271      43.000           21.500   \n",
       "25%        1.000     1.000    40.728    -73.992    2500.000         1225.000   \n",
       "50%        1.000     1.000    40.752    -73.978    3150.000         1500.000   \n",
       "75%        1.000     2.000    40.774    -73.955    4100.000         1850.000   \n",
       "max       10.000     8.000    44.883      0.000 4490000.000      2245000.000   \n",
       "\n",
       "       price_bedrooms  room_diff  room_num      Year       ...        \\\n",
       "count       49352.000  49352.000 49352.000 49352.000       ...         \n",
       "mean         1657.567     -0.329     2.754  2016.000       ...         \n",
       "std          7817.996      0.948     1.446     0.000       ...         \n",
       "min            43.000     -5.000     0.000  2016.000       ...         \n",
       "25%          1066.667     -1.000     2.000  2016.000       ...         \n",
       "50%          1383.417      0.000     2.000  2016.000       ...         \n",
       "75%          1962.500      0.000     4.000  2016.000       ...         \n",
       "max       1496666.667      8.000    13.500  2016.000       ...         \n",
       "\n",
       "           walk     walls       war    washer     water  wheelchair      wifi  \\\n",
       "count 49352.000 49352.000 49352.000 49352.000 49352.000   49352.000 49352.000   \n",
       "mean      0.003     0.000     0.186     0.009     0.000       0.028     0.002   \n",
       "std       0.055     0.020     0.389     0.102     0.021       0.165     0.045   \n",
       "min       0.000     0.000     0.000     0.000     0.000       0.000     0.000   \n",
       "25%       0.000     0.000     0.000     0.000     0.000       0.000     0.000   \n",
       "50%       0.000     0.000     0.000     0.000     0.000       0.000     0.000   \n",
       "75%       0.000     0.000     0.000     0.000     0.000       0.000     0.000   \n",
       "max       1.000     1.000     1.000     2.000     1.000       1.000     1.000   \n",
       "\n",
       "        windows      work  interest_level  \n",
       "count 49352.000 49352.000       49352.000  \n",
       "mean      0.001     0.001           1.617  \n",
       "std       0.032     0.031           0.626  \n",
       "min       0.000     0.000           0.000  \n",
       "25%       0.000     0.000           1.000  \n",
       "50%       0.000     0.000           2.000  \n",
       "75%       0.000     0.000           2.000  \n",
       "max       1.000     1.000           2.000  \n",
       "\n",
       "[8 rows x 225 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'Number of occurrences')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAELCAYAAAARNxsIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nREQ0JkUlIiIak6ISERGN+f+UQGoL+CmGNwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1131947f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Target 分布, 看看各类样本分布是否均衡\n",
    "sns.countplot(train.interest_level)\n",
    "pyplot.xlabel('interest_level')\n",
    "pyplot.ylabel('Number of occurrences')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "各类样本很不均衡. 交叉验证对分类任务缺省的是采用 StratifiedKFold，在每折采样时根据各类样本按比例采样."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>listing_id</th>\n",
       "      <th>bathrooms</th>\n",
       "      <th>bedrooms</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>price</th>\n",
       "      <th>price_bathrooms</th>\n",
       "      <th>price_bedrooms</th>\n",
       "      <th>room_diff</th>\n",
       "      <th>room_num</th>\n",
       "      <th>...</th>\n",
       "      <th>virtual</th>\n",
       "      <th>walk</th>\n",
       "      <th>walls</th>\n",
       "      <th>war</th>\n",
       "      <th>washer</th>\n",
       "      <th>water</th>\n",
       "      <th>wheelchair</th>\n",
       "      <th>wifi</th>\n",
       "      <th>windows</th>\n",
       "      <th>work</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7142618</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1</td>\n",
       "      <td>40.718</td>\n",
       "      <td>-73.987</td>\n",
       "      <td>2950</td>\n",
       "      <td>1475.000</td>\n",
       "      <td>1475.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>...</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>1</th>\n",
       "      <td>7210040</td>\n",
       "      <td>1.000</td>\n",
       "      <td>2</td>\n",
       "      <td>40.728</td>\n",
       "      <td>-74.000</td>\n",
       "      <td>2850</td>\n",
       "      <td>1425.000</td>\n",
       "      <td>950.000</td>\n",
       "      <td>-1.000</td>\n",
       "      <td>3.000</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7103890</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1</td>\n",
       "      <td>40.731</td>\n",
       "      <td>-73.989</td>\n",
       "      <td>3758</td>\n",
       "      <td>1879.000</td>\n",
       "      <td>1879.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>2.000</td>\n",
       "      <td>...</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>7143442</td>\n",
       "      <td>1.000</td>\n",
       "      <td>2</td>\n",
       "      <td>40.711</td>\n",
       "      <td>-73.957</td>\n",
       "      <td>3300</td>\n",
       "      <td>1650.000</td>\n",
       "      <td>1100.000</td>\n",
       "      <td>-1.000</td>\n",
       "      <td>3.000</td>\n",
       "      <td>...</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6860601</td>\n",
       "      <td>2.000</td>\n",
       "      <td>2</td>\n",
       "      <td>40.765</td>\n",
       "      <td>-73.984</td>\n",
       "      <td>4900</td>\n",
       "      <td>1633.333</td>\n",
       "      <td>1633.333</td>\n",
       "      <td>0.000</td>\n",
       "      <td>4.000</td>\n",
       "      <td>...</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 225 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   listing_id  bathrooms  bedrooms  latitude  longitude  price  \\\n",
       "0     7142618      1.000         1    40.718    -73.987   2950   \n",
       "1     7210040      1.000         2    40.728    -74.000   2850   \n",
       "2     7103890      1.000         1    40.731    -73.989   3758   \n",
       "3     7143442      1.000         2    40.711    -73.957   3300   \n",
       "4     6860601      2.000         2    40.765    -73.984   4900   \n",
       "\n",
       "   price_bathrooms  price_bedrooms  room_diff  room_num  ...   virtual  walk  \\\n",
       "0         1475.000        1475.000      0.000     2.000  ...         0     0   \n",
       "1         1425.000         950.000     -1.000     3.000  ...         0     0   \n",
       "2         1879.000        1879.000      0.000     2.000  ...         0     0   \n",
       "3         1650.000        1100.000     -1.000     3.000  ...         0     0   \n",
       "4         1633.333        1633.333      0.000     4.000  ...         0     0   \n",
       "\n",
       "   walls  war  washer  water  wheelchair  wifi  windows  work  \n",
       "0      0    0       0      0           0     0        0     0  \n",
       "1      0    1       0      0           0     0        0     0  \n",
       "2      0    0       0      0           0     0        0     0  \n",
       "3      0    0       0      0           1     0        0     0  \n",
       "4      0    1       0      0           0     0        0     0  \n",
       "\n",
       "[5 rows x 225 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 74659 entries, 0 to 74658\n",
      "Columns: 225 entries, listing_id to work\n",
      "dtypes: float64(7), int64(218)\n",
      "memory usage: 128.2 MB\n"
     ]
    }
   ],
   "source": [
    "test.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 数据预处理"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 提取 y, 在训练数据 X 中丢弃\n",
    "y_train = train['interest_level']\n",
    "train.drop(['interest_level'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = np.array(train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test_id = test['listing_id']\n",
    "test.drop(['listing_id'], axis=1, inplace=True)\n",
    "X_test = np.array(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 数据标准化\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "# 初始化特征的标准化器\n",
    "ss_X = StandardScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 分别对训练数据和测试数据的特征进行标准化处理\n",
    "X_train = ss_X.fit_transform(X_train)\n",
    "X_test = ss_X.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 模型训练"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## default Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lr = LogisticRegression()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/base.py:340: RuntimeWarning: overflow encountered in exp\n",
      "  np.exp(prob, prob)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/base.py:340: RuntimeWarning: overflow encountered in exp\n",
      "  np.exp(prob, prob)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/base.py:340: RuntimeWarning: overflow encountered in exp\n",
      "  np.exp(prob, prob)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "logloss of each fold is:  [ 0.68613212  0.69391517  0.68490632  0.67816236  0.6860222   0.6742054\n",
      "  0.6685211   0.66898296  0.67455859  0.6898933 ]\n",
      "cv loss is:  0.680529951061\n"
     ]
    }
   ],
   "source": [
    "# 交叉验证用于评估模型性能和进行参数调优(模型选择)\n",
    "# 分类任务中交叉验证缺省是采用 StratifiedKFold (分层采样 K折)\n",
    "from sklearn.cross_validation import cross_val_score\n",
    "loss = cross_val_score(lr, X_train, y_train, cv=10, scoring='neg_log_loss')\n",
    "print('logloss of each fold is: ', -loss)\n",
    "print('cv loss is: ', -loss.mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 正则化的 Logistic Regression 及参数调优"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Logistic Regression 需要调优的超参数有: C (正则系数, 一般在 log 域(取 log 后的值), 均匀设置候选参数) 和 正则函数 penalty (L2/L1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "目标函数为: J = sum(logloss(f(xi), yi)) + C * penalty"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在 sklearn 框架下, 不同学习器的参数调优步骤相同: 设置候选参数集合, 调用GridSearchCV, 调用 fit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 需要调优的参数\n",
    "penalty = ['l1']\n",
    "Cs = [1, 10, 100, 1000]\n",
    "\n",
    "tuned_parameters = dict(penalty=penalty, C=Cs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "lr_penalty = LogisticRegression()\n",
    "grid = GridSearchCV(lr_penalty, tuned_parameters, cv=5, scoring='neg_log_loss', n_jobs=-1, pre_dispatch=4)\n",
    "grid.fit(X_train[0:3999], y_train[0:3999])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/utils/deprecation.py:122: FutureWarning: You are accessing a training score ('mean_train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/utils/deprecation.py:122: FutureWarning: You are accessing a training score ('split0_train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/utils/deprecation.py:122: FutureWarning: You are accessing a training score ('split1_train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/utils/deprecation.py:122: FutureWarning: You are accessing a training score ('split2_train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/utils/deprecation.py:122: FutureWarning: You are accessing a training score ('split3_train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/utils/deprecation.py:122: FutureWarning: You are accessing a training score ('split4_train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/utils/deprecation.py:122: FutureWarning: You are accessing a training score ('std_train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'mean_fit_time': array([   0.13627062,    0.1546874 ,    0.63902845,    4.90874128,\n",
       "          25.06104388,   83.68735189,  113.85484724]),\n",
       " 'mean_score_time': array([ 0.00514183,  0.00371094,  0.00197253,  0.00157633,  0.00157151,\n",
       "         0.0016223 ,  0.00151401]),\n",
       " 'mean_test_score': array([-1.04325143, -0.79034616, -0.74774857, -0.78564641, -0.92766134,\n",
       "        -0.99380778, -1.06761922]),\n",
       " 'mean_train_score': array([-1.04320516, -0.78791156, -0.69787457, -0.63475706, -0.62742918,\n",
       "        -0.62591596, -0.62581106]),\n",
       " 'param_C': masked_array(data = [0.001 0.01 0.1 1 10 100 1000],\n",
       "              mask = [False False False False False False False],\n",
       "        fill_value = ?),\n",
       " 'param_penalty': masked_array(data = ['l1' 'l1' 'l1' 'l1' 'l1' 'l1' 'l1'],\n",
       "              mask = [False False False False False False False],\n",
       "        fill_value = ?),\n",
       " 'params': [{'C': 0.001, 'penalty': 'l1'},\n",
       "  {'C': 0.01, 'penalty': 'l1'},\n",
       "  {'C': 0.1, 'penalty': 'l1'},\n",
       "  {'C': 1, 'penalty': 'l1'},\n",
       "  {'C': 10, 'penalty': 'l1'},\n",
       "  {'C': 100, 'penalty': 'l1'},\n",
       "  {'C': 1000, 'penalty': 'l1'}],\n",
       " 'rank_test_score': array([6, 3, 1, 2, 4, 5, 7], dtype=int32),\n",
       " 'split0_test_score': array([-1.04337592, -0.79459915, -0.76283973, -0.80463892, -1.06821867,\n",
       "        -1.12960579, -1.22014229]),\n",
       " 'split0_train_score': array([-1.04327436, -0.78519685, -0.69389424, -0.63066823, -0.62272885,\n",
       "        -0.62099329, -0.62085169]),\n",
       " 'split1_test_score': array([-1.04329657, -0.78899463, -0.74346387, -0.74233326, -0.85015757,\n",
       "        -0.93690634, -1.01266056]),\n",
       " 'split1_train_score': array([-1.04311773, -0.78954249, -0.69820932, -0.63412217, -0.62585889,\n",
       "        -0.62396714, -0.6238455 ]),\n",
       " 'split2_test_score': array([-1.04328805, -0.79045635, -0.75451557, -0.82405202, -0.92670996,\n",
       "        -0.97331109, -0.98851631]),\n",
       " 'split2_train_score': array([-1.04311887, -0.78751009, -0.6968963 , -0.6346607 , -0.62693396,\n",
       "        -0.62554119, -0.62544777]),\n",
       " 'split3_test_score': array([-1.04320981, -0.7881755 , -0.73220467, -0.78623345, -0.86967686,\n",
       "        -0.93134369, -1.03685625]),\n",
       " 'split3_train_score': array([-1.04320803, -0.78835561, -0.69979985, -0.63478463, -0.62870616,\n",
       "        -0.62752667, -0.62744505]),\n",
       " 'split4_test_score': array([-1.0430864 , -0.7894961 , -0.74567814, -0.77093299, -0.92329003,\n",
       "        -0.99762896, -1.07970668]),\n",
       " 'split4_train_score': array([-1.04330681, -0.78895277, -0.70057315, -0.63954959, -0.63291803,\n",
       "        -0.6315515 , -0.63146532]),\n",
       " 'std_fit_time': array([  8.26029117e-03,   2.36413393e-02,   3.04803528e-01,\n",
       "          1.11001808e+00,   4.67335484e+00,   3.58163070e+01,\n",
       "          2.81591393e+01]),\n",
       " 'std_score_time': array([  5.66918125e-04,   9.85357415e-04,   5.27818277e-04,\n",
       "          3.80122069e-05,   7.88767936e-05,   1.03093649e-04,\n",
       "          9.89779333e-05]),\n",
       " 'std_test_score': array([  9.78161808e-05,   2.25314442e-03,   1.03721286e-02,\n",
       "          2.80347189e-02,   7.63847808e-02,   7.21664983e-02,\n",
       "          8.20604761e-02]),\n",
       " 'std_train_score': array([  7.77426804e-05,   1.51473259e-03,   2.36224870e-03,\n",
       "          2.83196002e-03,   3.36340190e-03,   3.53623474e-03,\n",
       "          3.55615888e-03])}"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 查看完整结果\n",
    "grid.cv_results_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.747748572614\n",
      "{'C': 0.1, 'penalty': 'l1'}\n"
     ]
    }
   ],
   "source": [
    "# 最好模型\n",
    "print(-grid.best_score_)\n",
    "print(grid.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 用 LogisticRegressionCV 实现正则化的 Logistic Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### L1 正则"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegressionCV(Cs=[0.01, 0.1, 1, 10, 100, 1000], class_weight=None,\n",
       "           cv=5, dual=False, fit_intercept=True, intercept_scaling=1.0,\n",
       "           max_iter=100, multi_class='ovr', n_jobs=-1, penalty='l1',\n",
       "           random_state=None, refit=True, scoring='neg_log_loss',\n",
       "           solver='liblinear', tol=0.0001, verbose=0)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegressionCV\n",
    "\n",
    "Cs = [0.01, 0.1, 1, 10, 100, 1000]\n",
    "\n",
    "# 大量样本, 高维度, L1正则 --> 可选用 saga 优化求解器\n",
    "# LogisticRegressionCV 比 GridSearchCV 快\n",
    "lrcv_L1 = LogisticRegressionCV(Cs=Cs, cv=5, scoring='neg_log_loss', penalty='l1', solver='liblinear', multi_class='ovr', n_jobs=-1)\n",
    "lrcv_L1.fit(X_train[:1999], y_train[:1999])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEKCAYAAADjDHn2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0db9c278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_Cs = len(Cs)\n",
    "n_classes = 3\n",
    "scores = np.zeros((n_classes, n_Cs))\n",
    "\n",
    "for j in range(n_classes):\n",
    "    scores[j][:] = np.mean(lrcv_L1.scores_[j], axis=0)\n",
    "    \n",
    "mse_mean = -np.mean(scores, axis=0)\n",
    "pyplot.plot(np.log10(Cs), mse_mean.reshape(n_Cs, 1))\n",
    "pyplot.xlabel('log(C)')\n",
    "pyplot.ylabel('neg-logloss')\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在 C=0.1 处表现最好."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: array([[-0.30739414, -0.28364392, -0.27915919, -0.29464118, -0.31296153,\n",
       "         -0.48574176],\n",
       "        [-0.30612974, -0.27115269, -0.28761289, -0.31423977, -0.33162505,\n",
       "         -0.52722594],\n",
       "        [-0.30878242, -0.27717846, -0.28352296, -0.30642243, -0.34221781,\n",
       "         -0.36850052],\n",
       "        [-0.30322027, -0.27308499, -0.31401815, -0.37785794, -0.44574409,\n",
       "         -0.51907922],\n",
       "        [-0.30414139, -0.2743943 , -0.25219057, -0.25382794, -0.26525285,\n",
       "         -0.27950065]]),\n",
       " 1: array([[-0.56263676, -0.55145827, -0.54794839, -0.56823048, -0.58986857,\n",
       "         -0.90334223],\n",
       "        [-0.56262782, -0.54362896, -0.55340613, -0.66872012, -0.7789763 ,\n",
       "         -0.97090767],\n",
       "        [-0.56118474, -0.5411046 , -0.63200902, -0.77412222, -0.87146702,\n",
       "         -1.02115163],\n",
       "        [-0.56162646, -0.53778705, -0.54833964, -0.59154628, -0.673801  ,\n",
       "         -0.7362706 ],\n",
       "        [-0.56207383, -0.53717575, -0.54898827, -0.57827781, -0.62212403,\n",
       "         -0.67493198]]),\n",
       " 2: array([[-0.63222633, -0.59362401, -0.55898906, -0.58050198, -0.64196002,\n",
       "         -1.05462072],\n",
       "        [-0.63336013, -0.59331633, -0.5732661 , -0.6340699 , -0.69841542,\n",
       "         -0.82342649],\n",
       "        [-0.63163232, -0.59533351, -0.65678193, -0.76415603, -0.86690371,\n",
       "         -1.135702  ],\n",
       "        [-0.63511748, -0.58534549, -0.59161651, -0.66112152, -0.72237599,\n",
       "         -0.80950363],\n",
       "        [-0.63140878, -0.59319759, -0.60042551, -0.67804891, -0.74134199,\n",
       "         -0.90045072]])}"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lrcv_L1.scores_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### L2 正则"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegressionCV(Cs=[0.01, 0.1, 1, 10, 100, 1000], class_weight=None,\n",
       "           cv=5, dual=False, fit_intercept=True, intercept_scaling=1.0,\n",
       "           max_iter=100, multi_class='ovr', n_jobs=-1, penalty='l2',\n",
       "           random_state=None, refit=True, scoring='neg_log_loss',\n",
       "           solver='liblinear', tol=0.0001, verbose=0)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegressionCV\n",
    "Cs = [0.01, 0.1, 1, 10, 100, 1000]\n",
    "\n",
    "lrcv_L2 = LogisticRegressionCV(Cs=Cs, cv=5, scoring='neg_log_loss', penalty='l2', solver='liblinear', multi_class='ovr', n_jobs=-1)\n",
    "lrcv_L2.fit(X_train[:1999], y_train[:1999])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: array([[-0.34141407, -0.32568811, -0.33538109, -0.35786995, -0.39639544,\n",
       "         -0.42624654],\n",
       "        [-0.32604407, -0.33291798, -0.36817062, -0.40834087, -0.43985653,\n",
       "         -0.45672087],\n",
       "        [-0.33423629, -0.31691667, -0.32399729, -0.33742662, -0.36392013,\n",
       "         -0.393501  ],\n",
       "        [-0.31108145, -0.31213092, -0.35982114, -0.40991671, -0.47306333,\n",
       "         -0.54634194],\n",
       "        [-0.34488117, -0.34633731, -0.35270074, -0.35330143, -0.36975914,\n",
       "         -0.3805916 ]]),\n",
       " 1: array([[-0.5766558 , -0.60636217, -0.62910475, -0.65489315, -0.68655325,\n",
       "         -0.71756114],\n",
       "        [-0.55280668, -0.57710438, -0.62731959, -0.70295126, -0.7954464 ,\n",
       "         -0.87800599],\n",
       "        [-0.57274751, -0.6200455 , -0.6911726 , -0.77159635, -0.85945032,\n",
       "         -0.94087032],\n",
       "        [-0.56997129, -0.58655692, -0.61716525, -0.66670463, -0.72084375,\n",
       "         -0.76975543],\n",
       "        [-0.57999249, -0.590667  , -0.60229849, -0.62883631, -0.64457982,\n",
       "         -0.65708907]]),\n",
       " 2: array([[-0.61219031, -0.6249594 , -0.62769831, -0.65178109, -0.6822253 ,\n",
       "         -0.70661991],\n",
       "        [-0.59351338, -0.61412426, -0.62863448, -0.66005651, -0.69938503,\n",
       "         -0.73946697],\n",
       "        [-0.63043471, -0.69084276, -0.75283273, -0.82740566, -0.91865395,\n",
       "         -1.01083987],\n",
       "        [-0.61356372, -0.64554461, -0.68676976, -0.7551689 , -0.8308737 ,\n",
       "         -0.91672585],\n",
       "        [-0.62949044, -0.6530497 , -0.67069267, -0.72340009, -0.78591738,\n",
       "         -0.82644499]])}"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lrcv_L2.scores_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEKCAYAAAA4t9PUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n5P8DYQrOHd1i3KEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a18a0e160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_Cs = len(Cs)\n",
    "n_classes = 3\n",
    "scores = np.zeros((n_classes, n_Cs))\n",
    "\n",
    "for j in range(n_classes):\n",
    "    scores[j][:] = np.mean(lrcv_L2.scores_[j], axis=0)\n",
    "    \n",
    "mse_mean = -np.mean(scores, axis=0)\n",
    "pyplot.plot(np.log10(Cs), mse_mean.reshape(n_Cs, 1))\n",
    "pyplot.xlabel('log(C)')\n",
    "pyplot.ylabel('neg-logloss')\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegressionCV(Cs=[1e-05, 0.0001, 0.001, 0.01], class_weight=None, cv=5,\n",
       "           dual=False, fit_intercept=True, intercept_scaling=1.0,\n",
       "           max_iter=100, multi_class='ovr', n_jobs=-1, penalty='l2',\n",
       "           random_state=None, refit=True, scoring='neg_log_loss',\n",
       "           solver='liblinear', tol=0.0001, verbose=0)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Cs = [0.00001, 0.0001, 0.001, 0.01]\n",
    "lrcv_L2 = LogisticRegressionCV(Cs=Cs, cv=5, scoring='neg_log_loss', penalty='l2', solver='liblinear', multi_class='ovr', n_jobs=-1)\n",
    "lrcv_L2.fit(X_train[:1999], y_train[:1999])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: array([[-0.6852168 , -0.62707519, -0.44728066, -0.34141407],\n",
       "        [-0.6850385 , -0.62511669, -0.43588938, -0.32604407],\n",
       "        [-0.68538714, -0.62822507, -0.44923032, -0.33423629],\n",
       "        [-0.68484355, -0.62366225, -0.43170686, -0.31108145],\n",
       "        [-0.68495513, -0.62520551, -0.44473002, -0.34488117]]),\n",
       " 1: array([[-0.68984729, -0.66589642, -0.59656899, -0.5766558 ],\n",
       "        [-0.68984954, -0.66565797, -0.58982898, -0.55280668],\n",
       "        [-0.68981502, -0.66571139, -0.59594069, -0.57274751],\n",
       "        [-0.68994604, -0.66674249, -0.59901216, -0.56997129],\n",
       "        [-0.69008904, -0.66796525, -0.60492585, -0.57999249]]),\n",
       " 2: array([[-0.69106554, -0.6758981 , -0.63020163, -0.61219031],\n",
       "        [-0.69080749, -0.67362157, -0.61788609, -0.59351338],\n",
       "        [-0.69102916, -0.67580232, -0.63274881, -0.63043471],\n",
       "        [-0.69098213, -0.67530611, -0.62861152, -0.61356372],\n",
       "        [-0.69116235, -0.67683733, -0.63580176, -0.62949044]])}"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lrcv_L2.scores_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEKCAYAAAA4t9PUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXhQkIiJSLwoSERGpl/8PEhZoH93PxI8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a19693be0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_Cs = len(Cs)\n",
    "n_classes = 3\n",
    "scores = np.zeros((n_classes, n_Cs))\n",
    "for j in range(n_classes):\n",
    "    scores[j][:] = np.mean(lrcv_L2.scores_[j], axis=0)\n",
    "    \n",
    "mse_mean = -np.mean(scores, axis=0)\n",
    "pyplot.plot(np.log10(Cs), mse_mean.reshape(n_Cs, 1))\n",
    "pyplot.xlabel('log(C)')\n",
    "pyplot.ylabel('neg-logloss')\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在 C=0.01 处表现最好."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 模型预测"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "LogisticRegressionCV在参数范围边缘(C=0.1)处取得极值. penalty 取'l1'表现更好."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/base.py:340: RuntimeWarning: overflow encountered in exp\n",
      "  np.exp(prob, prob)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/base.py:340: RuntimeWarning: overflow encountered in exp\n",
      "  np.exp(prob, prob)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/base.py:340: RuntimeWarning: overflow encountered in exp\n",
      "  np.exp(prob, prob)\n",
      "/Users/qi/anaconda3/lib/python3.6/site-packages/sklearn/linear_model/base.py:340: RuntimeWarning: overflow encountered in exp\n",
      "  np.exp(prob, prob)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegressionCV(Cs=[0.1], class_weight=None, cv=5, dual=False,\n",
       "           fit_intercept=True, intercept_scaling=1.0, max_iter=100,\n",
       "           multi_class='ovr', n_jobs=-1, penalty='l1', random_state=None,\n",
       "           refit=True, scoring='neg_log_loss', solver='liblinear',\n",
       "           tol=0.0001, verbose=0)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegressionCV\n",
    "Cs = [0.1]\n",
    "lrcv = LogisticRegressionCV(Cs=Cs, cv=5, scoring='neg_log_loss', penalty='l1', solver='liblinear', multi_class='ovr', n_jobs=-1)\n",
    "lrcv.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_predict = lrcv.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'Number of occurrences')"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAELCAYAAAARNxsIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEhERjfn/6Rp1ZsqFuL8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a0e6b5eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(y_predict)\n",
    "pyplot.xlabel('interest_level')\n",
    "pyplot.ylabel('Number of occurrences')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "直观上, 预测结果中 0 类(high)和 1 类(medium)比例太低."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "result = pd.DataFrame({'listing_id':test_id, 'interest_level_predict':y_predict})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "result.to_csv('LRresults_RentalListingInquiries.csv',index=False)"
   ]
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

一键复制编辑原始数据按行查看历史

提交于 2017-12-24 17:32 . Upload LR_RentalListingInquiries.ipynb

马建仓 AI 助手

尝试更多

代码解读

代码找茬

代码优化

Python

https://gitee.com/closeUrEyes/RentalListingInquiries.git

git@gitee.com:closeUrEyes/RentalListingInquiries.git

closeUrEyes

RentalListingInquiries

w2 - 亓呈文 - 854510525

master

Patrick Qi/w2 - 亓呈文 - 854510525

简介

发行版

贡献者

近期动态

Patrick Qi/w2 - 亓呈文 - 854510525 .gitee-modal { width: 500px !important; }

简介

发行版

贡献者

近期动态

搜索帮助

Patrick Qi/w2 - 亓呈文 - 854510525