DM_Airbnb/Untitled.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* [github](https://github.com/troywu666/Airbnb_data_analyse)\n",
    "* [blog](https://blog.csdn.net/datawhale/article/details/80847662)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn as sk\n",
    "%matplotlib inline\n",
    "import datetime\n",
    "import os\n",
    "import seaborn as sns#数据可视化\n",
    "from datetime import date\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "import pickle #用于存储模型\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import *\n",
    "from sklearn.model_selection import *"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "导入数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train=pd.read_csv(\"./input/train_users_2.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test=pd.read_csv(\"./input/test_users.csv\")"
   ]
  },
  {
   "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>id</th>\n",
       "      <th>date_account_created</th>\n",
       "      <th>timestamp_first_active</th>\n",
       "      <th>date_first_booking</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>signup_method</th>\n",
       "      <th>signup_flow</th>\n",
       "      <th>language</th>\n",
       "      <th>affiliate_channel</th>\n",
       "      <th>affiliate_provider</th>\n",
       "      <th>first_affiliate_tracked</th>\n",
       "      <th>signup_app</th>\n",
       "      <th>first_device_type</th>\n",
       "      <th>first_browser</th>\n",
       "      <th>country_destination</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>gxn3p5htnn</td>\n",
       "      <td>2010-06-28</td>\n",
       "      <td>20090319043255</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-unknown-</td>\n",
       "      <td>NaN</td>\n",
       "      <td>facebook</td>\n",
       "      <td>0</td>\n",
       "      <td>en</td>\n",
       "      <td>direct</td>\n",
       "      <td>direct</td>\n",
       "      <td>untracked</td>\n",
       "      <td>Web</td>\n",
       "      <td>Mac Desktop</td>\n",
       "      <td>Chrome</td>\n",
       "      <td>NDF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>820tgsjxq7</td>\n",
       "      <td>2011-05-25</td>\n",
       "      <td>20090523174809</td>\n",
       "      <td>NaN</td>\n",
       "      <td>MALE</td>\n",
       "      <td>38.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>0</td>\n",
       "      <td>en</td>\n",
       "      <td>seo</td>\n",
       "      <td>google</td>\n",
       "      <td>untracked</td>\n",
       "      <td>Web</td>\n",
       "      <td>Mac Desktop</td>\n",
       "      <td>Chrome</td>\n",
       "      <td>NDF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4ft3gnwmtx</td>\n",
       "      <td>2010-09-28</td>\n",
       "      <td>20090609231247</td>\n",
       "      <td>2010-08-02</td>\n",
       "      <td>FEMALE</td>\n",
       "      <td>56.0</td>\n",
       "      <td>basic</td>\n",
       "      <td>3</td>\n",
       "      <td>en</td>\n",
       "      <td>direct</td>\n",
       "      <td>direct</td>\n",
       "      <td>untracked</td>\n",
       "      <td>Web</td>\n",
       "      <td>Windows Desktop</td>\n",
       "      <td>IE</td>\n",
       "      <td>US</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bjjt8pjhuk</td>\n",
       "      <td>2011-12-05</td>\n",
       "      <td>20091031060129</td>\n",
       "      <td>2012-09-08</td>\n",
       "      <td>FEMALE</td>\n",
       "      <td>42.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>0</td>\n",
       "      <td>en</td>\n",
       "      <td>direct</td>\n",
       "      <td>direct</td>\n",
       "      <td>untracked</td>\n",
       "      <td>Web</td>\n",
       "      <td>Mac Desktop</td>\n",
       "      <td>Firefox</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>87mebub9p4</td>\n",
       "      <td>2010-09-14</td>\n",
       "      <td>20091208061105</td>\n",
       "      <td>2010-02-18</td>\n",
       "      <td>-unknown-</td>\n",
       "      <td>41.0</td>\n",
       "      <td>basic</td>\n",
       "      <td>0</td>\n",
       "      <td>en</td>\n",
       "      <td>direct</td>\n",
       "      <td>direct</td>\n",
       "      <td>untracked</td>\n",
       "      <td>Web</td>\n",
       "      <td>Mac Desktop</td>\n",
       "      <td>Chrome</td>\n",
       "      <td>US</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id date_account_created  timestamp_first_active date_first_booking  \\\n",
       "0  gxn3p5htnn           2010-06-28          20090319043255                NaN   \n",
       "1  820tgsjxq7           2011-05-25          20090523174809                NaN   \n",
       "2  4ft3gnwmtx           2010-09-28          20090609231247         2010-08-02   \n",
       "3  bjjt8pjhuk           2011-12-05          20091031060129         2012-09-08   \n",
       "4  87mebub9p4           2010-09-14          20091208061105         2010-02-18   \n",
       "\n",
       "      gender   age signup_method  signup_flow language affiliate_channel  \\\n",
       "0  -unknown-   NaN      facebook            0       en            direct   \n",
       "1       MALE  38.0      facebook            0       en               seo   \n",
       "2     FEMALE  56.0         basic            3       en            direct   \n",
       "3     FEMALE  42.0      facebook            0       en            direct   \n",
       "4  -unknown-  41.0         basic            0       en            direct   \n",
       "\n",
       "  affiliate_provider first_affiliate_tracked signup_app first_device_type  \\\n",
       "0             direct               untracked        Web       Mac Desktop   \n",
       "1             google               untracked        Web       Mac Desktop   \n",
       "2             direct               untracked        Web   Windows Desktop   \n",
       "3             direct               untracked        Web       Mac Desktop   \n",
       "4             direct               untracked        Web       Mac Desktop   \n",
       "\n",
       "  first_browser country_destination  \n",
       "0        Chrome                 NDF  \n",
       "1        Chrome                 NDF  \n",
       "2            IE                  US  \n",
       "3       Firefox               other  \n",
       "4        Chrome                  US  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- id: user id （用户id）\n",
    "- date_account_created（帐号注册时间）: the date of account creation\n",
    "- timestamp_first_active（首次活跃时间）: timestamp of the first activity, note that it can be earlier than date_account_created or date_first_booking because a user can search before signing up\n",
    "- date_first_booking（首次订房时间）: date of first booking\n",
    "- gender（性别）\n",
    "- age（年龄）\n",
    "- signup_method（注册方式）\n",
    "- signup_flow（注册页面）: the page a user came to signup up from\n",
    "- language（语言）: international language preference\n",
    "- affiliate_channel（付费市场渠道）: what kind of paid marketing\n",
    "- affiliate_provider（付费市场渠道名称）: where the marketing is e.g. google, craigslist, other\n",
    "- first_affiliate_tracked（注册前第一个接触的市场渠道）: whats the first marketing the user interacted with before the signing up\n",
    "- signup_app（注册app）\n",
    "- first_device_type(设备类型)\n",
    "- first_browser（浏览器类型）\n",
    "- country_destination（订房国家-需要预测的量）: this is the target variable you are to predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(213451, 16)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(62096, 15)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'country_destination'}"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set(train.columns)-set(test.columns)   #country_destination是需要预测的目标变量"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 213451 entries, 0 to 213450\n",
      "Data columns (total 16 columns):\n",
      "id                         213451 non-null object\n",
      "date_account_created       213451 non-null object\n",
      "timestamp_first_active     213451 non-null int64\n",
      "date_first_booking         88908 non-null object\n",
      "gender                     213451 non-null object\n",
      "age                        125461 non-null float64\n",
      "signup_method              213451 non-null object\n",
      "signup_flow                213451 non-null int64\n",
      "language                   213451 non-null object\n",
      "affiliate_channel          213451 non-null object\n",
      "affiliate_provider         213451 non-null object\n",
      "first_affiliate_tracked    207386 non-null object\n",
      "signup_app                 213451 non-null object\n",
      "first_device_type          213451 non-null object\n",
      "first_browser              213451 non-null object\n",
      "country_destination        213451 non-null object\n",
      "dtypes: float64(1), int64(2), object(13)\n",
      "memory usage: 26.1+ MB\n"
     ]
    }
   ],
   "source": [
    "train.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#文件包含213451行数据，16个特征，每个特征的数据类型和非空数值，date_first_booking空值较多"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 对每个字段逐一分析"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. date_account_created"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count         213451\n",
       "unique          1634\n",
       "top       2014-05-13\n",
       "freq             674\n",
       "Name: date_account_created, dtype: object"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.date_account_created.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 统计并查看分布"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2014-05-13    674\n",
      "2014-06-24    670\n",
      "2014-06-25    636\n",
      "2014-05-20    632\n",
      "2014-05-14    622\n",
      "Name: date_account_created, dtype: int64\n",
      "2010-01-28    1\n",
      "2010-01-19    1\n",
      "2010-04-11    1\n",
      "2010-04-24    1\n",
      "2010-06-18    1\n",
      "Name: date_account_created, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(train.date_account_created.value_counts().head())\n",
    "print(train.date_account_created.value_counts().tail())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c54ff14e0>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c54ff1278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dac_train = train.date_account_created.value_counts()\n",
    "dac_test = test.date_account_created.value_counts()\n",
    "#将数据类型转换为datatime类型\n",
    "dac_train_date = pd.to_datetime(train.date_account_created.value_counts().index)\n",
    "dac_test_date = pd.to_datetime(test.date_account_created.value_counts().index)\n",
    "#计算离首次注册时间相差的天数\n",
    "dac_train_day = dac_train_date - dac_train_date.min()\n",
    "dac_test_day = dac_test_date - dac_train_date.min()\n",
    "#motplotlib作图\n",
    "plt.scatter(dac_train_day.days, dac_train.values, color = 'r', label = 'train dataset')\n",
    "plt.scatter(dac_test_day.days, dac_test.values, color = 'b', label = 'test dataset')\n",
    "\n",
    "plt.title(\"Accounts created vs day\")\n",
    "plt.xlabel(\"Days\")\n",
    "plt.ylabel(\"Accounts created\")\n",
    "plt.legend(loc = 'upper center')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. x轴表示离首次注册的天数，y轴表示当天注册的用户数量\n",
    "2. 随着时间的增长,用户注册的数量在急剧上升"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. timestamp_first_active"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    20090319043255\n",
       "1    20090523174809\n",
       "2    20090609231247\n",
       "3    20091031060129\n",
       "4    20091208061105\n",
       "Name: timestamp_first_active, dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.timestamp_first_active.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "对数据进行统计看非重复值的数量"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1], dtype=int64)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.timestamp_first_active.value_counts().unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "结果[1]表明timestamp_first_active没有重复数据"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* 将object数据类型转换为时间数据类型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0   2009-03-19 04:32:55\n",
       "1   2009-05-23 17:48:09\n",
       "2   2009-06-09 23:12:47\n",
       "3   2009-10-31 06:01:29\n",
       "4   2009-12-08 06:11:05\n",
       "Name: timestamp_first_active, dtype: datetime64[ns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tfa_train=train.timestamp_first_active\n",
    "tfa_train=train.timestamp_first_active.map(lambda x: datetime.datetime.strptime(str(x),'%Y%m%d%H%M%S'))\n",
    "tfa_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 查看用户一天中活跃的时间段"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x20c593ce208>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAD8CAYAAACcjGjIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nB6WUUs1ocFBKKdWMBgellFLN/H99gPIW12h5KwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c5b7df6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.pointplot(x=tfa_train.dt.hour.value_counts().index,y=tfa_train.dt.hour.value_counts().values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "用户首次活跃时间在一天中集中分布在17点之后到凌晨5点"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. date_first_booking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count          88908\n",
       "unique          1976\n",
       "top       2014-05-22\n",
       "freq             248\n",
       "Name: date_first_booking, dtype: object"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.date_first_booking.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5834734904029496"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.date_first_booking.isnull().sum()/213451"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "首次订房时间缺失高达50%以上，可考虑删除该字段"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. age"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "30.0    6124\n",
       "31.0    6016\n",
       "29.0    5963\n",
       "28.0    5939\n",
       "32.0    5855\n",
       "27.0    5738\n",
       "33.0    5527\n",
       "26.0    5044\n",
       "34.0    5029\n",
       "35.0    4860\n",
       "Name: age, dtype: int64"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.age.value_counts(ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "用户年龄主要集中在30左右"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'counts')"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnoAAAE/CAYAAADR125OAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUfBg7wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c55aa5278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#首先将年龄进行分成4组missing values, too small age, reasonable age, too large age\n",
    "age_train =[train[train.age.isnull()].age.shape[0],\n",
    "            train.query('age < 15').age.shape[0],\n",
    "            train.query(\"age >= 15 & age <= 90\").age.shape[0],\n",
    "            train.query('age > 90').age.shape[0]]\n",
    "\n",
    "age_test = [test[test.age.isnull()].age.shape[0],\n",
    "            test.query('age < 15').age.shape[0],\n",
    "            test.query(\"age >= 15 & age <= 90\").age.shape[0],\n",
    "            test.query('age > 90').age.shape[0]]\n",
    "\n",
    "columns = ['Null', 'age < 15', 'age', 'age > 90']\n",
    "\n",
    "# plot\n",
    "fig, (ax1,ax2) = plt.subplots(1,2,sharex=True, sharey = True,figsize=(10,5))\n",
    "\n",
    "sns.barplot(columns, age_train, ax = ax1)\n",
    "sns.barplot(columns, age_test, ax = ax2)\n",
    "\n",
    "ax1.set_title('training dataset')\n",
    "ax2.set_title('test dataset')\n",
    "ax1.set_ylabel('counts')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#统一使用柱状图进行统计\n",
    "def feature_barplot(feature, df_train = train, df_test = test, figsize=(10,5), rot = 90, saveimg = False): \n",
    "    feat_train = df_train[feature].value_counts()\n",
    "    feat_test = df_test[feature].value_counts()\n",
    "    fig_feature, (axis1,axis2) = plt.subplots(1,2,sharex=True, sharey = True, figsize = figsize)\n",
    "    sns.barplot(feat_train.index.values, feat_train.values, ax = axis1)\n",
    "    sns.barplot(feat_test.index.values, feat_test.values, ax = axis2)\n",
    "    axis1.set_xticklabels(axis1.xaxis.get_majorticklabels(), rotation = rot)\n",
    "    axis2.set_xticklabels(axis1.xaxis.get_majorticklabels(), rotation = rot)\n",
    "    axis1.set_title(feature + ' of training dataset')\n",
    "    axis2.set_title(feature + ' of test dataset')\n",
    "    axis1.set_ylabel('Counts')\n",
    "    plt.tight_layout()\n",
    "    if saveimg == True:\n",
    "        figname = feature + \".png\"\n",
    "        fig_feature.savefig(figname, dpi = 75)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAAFgCAYAAACmDI9oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c5475f710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_barplot('gender', saveimg = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAAFgCAYAAACmDI9oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJEmSeiyQJUmSpB4LZEmSJKnn/wPzvLsT5A+RBQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c5cca6da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_barplot('signup_method')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAAFgCAYAAACmDI9oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n6rFAliRJknoskCVJkqSe/x/wxR93J1yBdgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c5c6946a0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_barplot('signup_app')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# sesion文件"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "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>user_id</th>\n",
       "      <th>action</th>\n",
       "      <th>action_type</th>\n",
       "      <th>action_detail</th>\n",
       "      <th>device_type</th>\n",
       "      <th>secs_elapsed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>d1mm9tcy42</td>\n",
       "      <td>lookup</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Windows Desktop</td>\n",
       "      <td>319.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>d1mm9tcy42</td>\n",
       "      <td>search_results</td>\n",
       "      <td>click</td>\n",
       "      <td>view_search_results</td>\n",
       "      <td>Windows Desktop</td>\n",
       "      <td>67753.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>d1mm9tcy42</td>\n",
       "      <td>lookup</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Windows Desktop</td>\n",
       "      <td>301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>d1mm9tcy42</td>\n",
       "      <td>search_results</td>\n",
       "      <td>click</td>\n",
       "      <td>view_search_results</td>\n",
       "      <td>Windows Desktop</td>\n",
       "      <td>22141.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>jpmpj40cvk</td>\n",
       "      <td>lookup</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Windows Desktop</td>\n",
       "      <td>957.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      user_id          action action_type        action_detail  \\\n",
       "0  d1mm9tcy42          lookup         NaN                  NaN   \n",
       "1  d1mm9tcy42  search_results       click  view_search_results   \n",
       "2  d1mm9tcy42          lookup         NaN                  NaN   \n",
       "3  d1mm9tcy42  search_results       click  view_search_results   \n",
       "4  jpmpj40cvk          lookup         NaN                  NaN   \n",
       "\n",
       "       device_type  secs_elapsed  \n",
       "0  Windows Desktop         319.0  \n",
       "1  Windows Desktop       67753.0  \n",
       "2  Windows Desktop         301.0  \n",
       "3  Windows Desktop       22141.0  \n",
       "4  Windows Desktop         957.0  "
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sessions=pd.read_csv('./input/sessions.csv')\n",
    "df_sessions.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#将user_id改名为id\n",
    "df_sessions.rename(columns={'user_id':'id'},inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1048575, 6)"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sessions.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "查看缺失值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                 2797\n",
       "action             7873\n",
       "action_type      143097\n",
       "action_detail    143097\n",
       "device_type           0\n",
       "secs_elapsed      11067\n",
       "dtype: int64"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sessions.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. action，action_type，action_detail， secs_elapsed缺失值较多\n",
    "2. 填充缺失值，(此处用了字符串'NAN')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                2797\n",
       "action               0\n",
       "action_type          0\n",
       "action_detail        0\n",
       "device_type          0\n",
       "secs_elapsed     11067\n",
       "dtype: int64"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sessions.action = df_sessions.action.fillna('NAN')\n",
    "df_sessions.action_type = df_sessions.action_type.fillna('NAN')\n",
    "df_sessions.action_detail = df_sessions.action_detail.fillna('NAN')\n",
    "df_sessions.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#对action进行统计，我们可以发现用户action有多种，且最少的发生次数只有1，接下来我们可以对用户发生次数较少的行为列为OTHER一类"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "payoneer_account_redirect          9\n",
       "invalid_action                     9\n",
       "redirect                           9\n",
       "request_photography                8\n",
       "pricing                            7\n",
       "become_user                        7\n",
       "press_content                      7\n",
       "hospitality                        7\n",
       "apply_code                         6\n",
       "city_count                         6\n",
       "multi                              6\n",
       "onenight                           6\n",
       "social                             5\n",
       "ajax_send_message                  5\n",
       "create_paypal                      5\n",
       "apply                              5\n",
       "new_session                        5\n",
       "zendesk_login_jwt                  4\n",
       "phone_verification                 4\n",
       "google_importer                    4\n",
       "payoneer_signup_complete           4\n",
       "founders                           4\n",
       "change_availability                4\n",
       "ajax_photo_widget                  4\n",
       "respond                            4\n",
       "office_location                    3\n",
       "press_news                         3\n",
       "slideshow                          3\n",
       "locations                          3\n",
       "message                            3\n",
       "travel                             3\n",
       "locale_from_host                   2\n",
       "create_ach                         2\n",
       "home_safety_landing                2\n",
       "has_profile_pic                    2\n",
       "how_it_works                       2\n",
       "load_more                          2\n",
       "reservation                        2\n",
       "update_reservation_requirements    2\n",
       "life                               2\n",
       "media_resources                    1\n",
       "wishlists                          1\n",
       "maybe_information                  1\n",
       "feed                               1\n",
       "change_default_payout              1\n",
       "guest_billing_receipt              1\n",
       "localized                          1\n",
       "recommendation_page                1\n",
       "southern-europe                    1\n",
       "relationship                       1\n",
       "photography_update                 1\n",
       "satisfy                            1\n",
       "departments                        1\n",
       "sublets                            1\n",
       "Name: action, dtype: int64"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sessions.action.value_counts()[df_sessions.action.value_counts()<10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Action values with low frequency are changed to 'OTHER'\n",
    "act_freq = 100  #Threshold of frequency\n",
    "act = dict(zip(*np.unique(df_sessions.action, return_counts=True)))\n",
    "df_sessions.action = df_sessions.action.apply(lambda x: 'OTHER' if act[x] < act_freq else x)\n",
    "#np.unique(df_sessions.action, return_counts=True) 取以数组形式返回非重复的action值和它的数量\n",
    "#zip（*（a,b））a,b种元素一一对应，返回zip object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'10': 200,\n",
       " '12': 282,\n",
       " 'NAN': 7873,\n",
       " 'OTHER': 3184,\n",
       " 'account': 1112,\n",
       " 'active': 15855,\n",
       " 'add_note': 161,\n",
       " 'ajax_check_dates': 8387,\n",
       " 'ajax_get_referrals_amt': 557,\n",
       " 'ajax_image_upload': 1560,\n",
       " 'ajax_lwlb_contact': 5042,\n",
       " 'ajax_payout_edit': 103,\n",
       " 'ajax_payout_options_by_country': 115,\n",
       " 'ajax_photo_widget_form_iframe': 2424,\n",
       " 'ajax_refresh_subtotal': 71482,\n",
       " 'ajax_statsd': 307,\n",
       " 'apply_reservation': 1272,\n",
       " 'ask_question': 6994,\n",
       " 'at_checkpoint': 497,\n",
       " 'authenticate': 6021,\n",
       " 'authorize': 127,\n",
       " 'available': 717,\n",
       " 'calendar_tab_inner2': 9353,\n",
       " 'callback': 888,\n",
       " 'campaigns': 2943,\n",
       " 'cancellation_policies': 1592,\n",
       " 'click': 555,\n",
       " 'collections': 12763,\n",
       " 'complete': 320,\n",
       " 'complete_status': 1790,\n",
       " 'confirm_email': 7424,\n",
       " 'connect': 1000,\n",
       " 'contact_new': 175,\n",
       " 'create': 17262,\n",
       " 'create_multiple': 153,\n",
       " 'dashboard': 9535,\n",
       " 'decision_tree': 308,\n",
       " 'delete': 623,\n",
       " 'domains': 438,\n",
       " 'edit': 7716,\n",
       " 'edit_verification': 2288,\n",
       " 'facebook_auto_login': 550,\n",
       " 'faq': 1192,\n",
       " 'faq_category': 906,\n",
       " 'faq_experiment_ids': 704,\n",
       " 'glob': 119,\n",
       " 'header_userpic': 17332,\n",
       " 'hosting_social_proof': 2510,\n",
       " 'identity': 2846,\n",
       " 'index': 69441,\n",
       " 'itinerary': 127,\n",
       " 'jumio_redirect': 446,\n",
       " 'jumio_token': 751,\n",
       " 'kba': 1012,\n",
       " 'kba_update': 1619,\n",
       " 'languages_multiselect': 662,\n",
       " 'listing': 136,\n",
       " 'listings': 4327,\n",
       " 'login': 276,\n",
       " 'login_modal': 585,\n",
       " 'lookup': 88302,\n",
       " 'manage_listing': 2144,\n",
       " 'my': 1710,\n",
       " 'new': 817,\n",
       " 'notifications': 3420,\n",
       " 'open_graph_setting': 1341,\n",
       " 'other_hosting_reviews': 142,\n",
       " 'other_hosting_reviews_first': 5934,\n",
       " 'pay': 104,\n",
       " 'payment_instruments': 1113,\n",
       " 'payment_methods': 414,\n",
       " 'payout_preferences': 337,\n",
       " 'pending': 2844,\n",
       " 'pending_tickets': 136,\n",
       " 'personalize': 105718,\n",
       " 'phone_number_widget': 875,\n",
       " 'phone_verification_modal': 227,\n",
       " 'popular': 155,\n",
       " 'populate_from_facebook': 746,\n",
       " 'populate_help_dropdown': 946,\n",
       " 'profile_pic': 2126,\n",
       " 'qt2': 4537,\n",
       " 'qt_reply_v2': 1147,\n",
       " 'qt_with': 194,\n",
       " 'recent_reservations': 272,\n",
       " 'references': 810,\n",
       " 'referrer_status': 189,\n",
       " 'remove_dashboard_alert': 144,\n",
       " 'request_new_confirm_email': 794,\n",
       " 'requested': 6242,\n",
       " 'reviews': 4704,\n",
       " 'reviews_new': 616,\n",
       " 'search': 26520,\n",
       " 'search_results': 109430,\n",
       " 'set_user': 937,\n",
       " 'settings': 447,\n",
       " 'show': 209655,\n",
       " 'show_personalize': 5542,\n",
       " 'signature': 327,\n",
       " 'signup_login': 395,\n",
       " 'signup_modal': 356,\n",
       " 'similar_listings': 53814,\n",
       " 'social_connections': 15301,\n",
       " 'status': 270,\n",
       " 'supported': 176,\n",
       " 'tell_a_friend': 247,\n",
       " 'this_hosting_reviews': 816,\n",
       " 'tos_confirm': 309,\n",
       " 'track_page_view': 17089,\n",
       " 'transaction_history': 113,\n",
       " 'transaction_history_paginated': 221,\n",
       " 'travel_plans_current': 2965,\n",
       " 'unavailabilities': 4702,\n",
       " 'update': 38829,\n",
       " 'update_notifications': 176,\n",
       " 'uptodate': 462,\n",
       " 'verify': 3337}"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict(zip(*np.unique(df_sessions.action,return_counts=True)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "首先将用户的特征根据用户id进行分组\n",
    "- **特征action：**统计每个用户总的action出现的次数，各个action类型的数量，平均值以及标准差\n",
    "- **特征action_detail：**统计每个用户总的action_detail出现的次数，各个action_detail类型的数量，平均值以及标准差\n",
    "- **特征action_type：**统计每个用户总的action_type出现的次数，各个action_type类型的数量，平均值，标准差以及总的停留时长（进行log处理）\n",
    "- **特征device_type：**统计每个用户总的device_type出现的次数，各个device_type类型的数量，平均值以及标准差\n",
    "- **特征secs_elapsed：**对缺失值用0填充，统计每个用户secs_elapsed时间的总和，平均值，标准差以及中位数（进行log处理），（总和/平均数），secs_elapsed（log处理后）各个时间出现的次数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#对action特征进行细化\n",
    "f_act = df_sessions.action.value_counts().argsort()\n",
    "f_act_detail = df_sessions.action_detail.value_counts().argsort()\n",
    "f_act_type = df_sessions.action_type.value_counts().argsort()\n",
    "f_dev_type = df_sessions.device_type.value_counts().argsort()\n",
    "\n",
    "#按照id进行分组\n",
    "dgr_sess = df_sessions.groupby(['id'])\n",
    "#Loop on dgr_sess to create all the features.\n",
    "samples = [] #samples列表\n",
    "ln = len(dgr_sess) #计算分组后df_sessions的长度\n",
    "\n",
    "for g in dgr_sess:  #对dgr_sess中每个id的数据进行遍历\n",
    "    gr = g[1]   #data frame that comtains all the data for a groupby value 'zzywmcn0jv'\n",
    "    \n",
    "    l = []  #建一个空列表，临时存放特征\n",
    "    \n",
    "    #the id    for example:'zzywmcn0jv'\n",
    "    l.append(g[0]) #将id值放入空列表中\n",
    "    \n",
    "    # number of total actions\n",
    "    l.append(len(gr))#将id对应数据的长度放入列表\n",
    "    \n",
    "    #secs_elapsed 特征中的缺失值用0填充再获取具体的停留时长值\n",
    "    sev = gr.secs_elapsed.fillna(0).values   #These values are used later.\n",
    "    \n",
    "    #action features 特征-用户行为 \n",
    "    #每个用户行为出现的次数，各个行为类型的数量，平均值以及标准差\n",
    "    c_act = [0] * len(f_act)\n",
    "    for i,v in enumerate(gr.action.values): #i是从0-1对应的位置，v 是用户行为特征的值\n",
    "        c_act[f_act[v]] += 1\n",
    "    _, c_act_uqc = np.unique(gr.action.values, return_counts=True)\n",
    "    #计算用户行为行为特征各个类型数量的长度，平均值以及标准差\n",
    "    c_act += [len(c_act_uqc), np.mean(c_act_uqc), np.std(c_act_uqc)]\n",
    "    l = l + c_act\n",
    "    \n",
    "    #action_detail features 特征-用户行为具体\n",
    "    #(how many times each value occurs, numb of unique values, mean and std)\n",
    "    c_act_detail = [0] * len(f_act_detail)\n",
    "    for i,v in enumerate(gr.action_detail.values):\n",
    "        c_act_detail[f_act_detail[v]] += 1\n",
    "    _, c_act_det_uqc = np.unique(gr.action_detail.values, return_counts=True)\n",
    "    c_act_detail += [len(c_act_det_uqc), np.mean(c_act_det_uqc), np.std(c_act_det_uqc)]\n",
    "    l = l + c_act_detail\n",
    "    \n",
    "    #action_type features  特征-用户行为类型 click等\n",
    "    #(how many times each value occurs, numb of unique values, mean and std\n",
    "    #+ log of the sum of secs_elapsed for each value)\n",
    "    l_act_type = [0] * len(f_act_type)\n",
    "    c_act_type = [0] * len(f_act_type)\n",
    "    for i,v in enumerate(gr.action_type.values):\n",
    "        l_act_type[f_act_type[v]] += sev[i] #sev = gr.secs_elapsed.fillna(0).values ，求每个行为类型总的停留时长\n",
    "        c_act_type[f_act_type[v]] += 1  \n",
    "    l_act_type = np.log(1 + np.array(l_act_type)).tolist() #每个行为类型总的停留时长，差异比较大，进行log处理\n",
    "    _, c_act_type_uqc = np.unique(gr.action_type.values, return_counts=True)\n",
    "    c_act_type += [len(c_act_type_uqc), np.mean(c_act_type_uqc), np.std(c_act_type_uqc)]\n",
    "    l = l + c_act_type + l_act_type    \n",
    "    \n",
    "    #device_type features 特征-设备类型\n",
    "    #(how many times each value occurs, numb of unique values, mean and std)\n",
    "    c_dev_type  = [0] * len(f_dev_type)\n",
    "    for i,v in enumerate(gr.device_type .values):\n",
    "        c_dev_type[f_dev_type[v]] += 1 \n",
    "    c_dev_type.append(len(np.unique(gr.device_type.values))) \n",
    "    _, c_dev_type_uqc = np.unique(gr.device_type.values, return_counts=True)\n",
    "    c_dev_type += [len(c_dev_type_uqc), np.mean(c_dev_type_uqc), np.std(c_dev_type_uqc)]        \n",
    "    l = l + c_dev_type    \n",
    "    \n",
    "    #secs_elapsed features  特征-停留时长     \n",
    "    l_secs = [0] * 5 \n",
    "    l_log = [0] * 15\n",
    "    if len(sev) > 0:\n",
    "        #Simple statistics about the secs_elapsed values.\n",
    "        l_secs[0] = np.log(1 + np.sum(sev))\n",
    "        l_secs[1] = np.log(1 + np.mean(sev)) \n",
    "        l_secs[2] = np.log(1 + np.std(sev))\n",
    "        l_secs[3] = np.log(1 + np.median(sev))\n",
    "        l_secs[4] = l_secs[0] / float(l[1]) #\n",
    "        \n",
    "        #Values are grouped in 15 intervals. Compute the number of values\n",
    "        #in each interval.\n",
    "        #sev = gr.secs_elapsed.fillna(0).values \n",
    "        log_sev = np.log(1 + sev).astype(int)\n",
    "        #np.bincount():Count number of occurrences of each value in array of non-negative ints.  \n",
    "        l_log = np.bincount(log_sev, minlength=15).tolist()                    \n",
    "    l = l + l_secs + l_log\n",
    "    \n",
    "    #The list l has the feature values of one sample.\n",
    "    samples.append(l)\n",
    "\n",
    "#preparing objects    \n",
    "samples = np.array(samples) \n",
    "samp_ar = samples[:, 1:].astype(np.float16) #取除id外的特征数据\n",
    "samp_id = samples[:, 0]   #取id，id位于第一列\n",
    "\n",
    "#为提取的特征创建一个dataframe     \n",
    "col_names = []    #name of the columns\n",
    "for i in range(len(samples[0])-1):  #减1的原因是因为有个id\n",
    "    col_names.append('c_' + str(i))  #起名字的方式    \n",
    "df_agg_sess = pd.DataFrame(samp_ar, columns=col_names)\n",
    "df_agg_sess['id'] = samp_id\n",
    "df_agg_sess.index = df_agg_sess.id #将id作为index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "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>c_0</th>\n",
       "      <th>c_1</th>\n",
       "      <th>c_2</th>\n",
       "      <th>c_3</th>\n",
       "      <th>c_4</th>\n",
       "      <th>c_5</th>\n",
       "      <th>c_6</th>\n",
       "      <th>c_7</th>\n",
       "      <th>c_8</th>\n",
       "      <th>c_9</th>\n",
       "      <th>...</th>\n",
       "      <th>c_284</th>\n",
       "      <th>c_285</th>\n",
       "      <th>c_286</th>\n",
       "      <th>c_287</th>\n",
       "      <th>c_288</th>\n",
       "      <th>c_289</th>\n",
       "      <th>c_290</th>\n",
       "      <th>c_291</th>\n",
       "      <th>c_292</th>\n",
       "      <th>id</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>zzqb2sn066</th>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>zzqb2sn066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>zzrnx9rqi5</th>\n",
       "      <td>26.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>zzrnx9rqi5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>zzv8sgicbk</th>\n",
       "      <td>25.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>zzv8sgicbk</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>zzvatt4dio</th>\n",
       "      <td>422.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>108.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>zzvatt4dio</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>zzywmcn0jv</th>\n",
       "      <td>51.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>zzywmcn0jv</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 294 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              c_0  c_1  c_2  c_3  c_4  c_5  c_6  c_7  c_8  c_9     ...      \\\n",
       "id                                                                 ...       \n",
       "zzqb2sn066   10.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0     ...       \n",
       "zzrnx9rqi5   26.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0     ...       \n",
       "zzv8sgicbk   25.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0     ...       \n",
       "zzvatt4dio  422.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0     ...       \n",
       "zzywmcn0jv   51.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0     ...       \n",
       "\n",
       "            c_284  c_285  c_286  c_287  c_288  c_289  c_290  c_291  c_292  \\\n",
       "id                                                                          \n",
       "zzqb2sn066    1.0    2.0    0.0    0.0    1.0    0.0    0.0    1.0    0.0   \n",
       "zzrnx9rqi5    9.0    3.0    1.0    1.0    5.0    0.0    1.0    1.0    0.0   \n",
       "zzv8sgicbk    5.0    3.0    4.0    3.0    3.0    2.0    1.0    0.0    0.0   \n",
       "zzvatt4dio  108.0   79.0   26.0   41.0   32.0   20.0    8.0    3.0    2.0   \n",
       "zzywmcn0jv    3.0   15.0    5.0    5.0    1.0    3.0    2.0    2.0    0.0   \n",
       "\n",
       "                    id  \n",
       "id                      \n",
       "zzqb2sn066  zzqb2sn066  \n",
       "zzrnx9rqi5  zzrnx9rqi5  \n",
       "zzv8sgicbk  zzv8sgicbk  \n",
       "zzvatt4dio  zzvatt4dio  \n",
       "zzywmcn0jv  zzywmcn0jv  \n",
       "\n",
       "[5 rows x 294 columns]"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_agg_sess.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15588, 294)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_agg_sess.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4.2 对trian和test文件进行特征提取"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#计算出train的行数，便于之后对train和test数据进行分离操作\n",
    "train_row = train.shape[0]  \n",
    "\n",
    "# The label we need to predict\n",
    "labels = train['country_destination'].values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 数据探索时我们发现date_first_booking在train和test文件中缺失值太多，故删除\n",
    "- 删除country_destination，用模型预测country_destination，再与已经存储country_destination的labels进行比较，从而判断模型优劣"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.drop(['country_destination', 'date_first_booking'], axis = 1, inplace = True)\n",
    "test.drop(['date_first_booking'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "合并train和test文件\n",
    "- 便于进行相同的特征提取操作"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#连接test 和 train\n",
    "df = pd.concat([train, test], axis = 0, ignore_index = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(275547, 14)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## timestamp_first_active \n",
    "1.1转换为datetime类型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tfa=df.timestamp_first_active.astype(str).apply(lambda x: datetime.datetime(int(x[:4]),\n",
    "                                                                          int(x[4:6]), \n",
    "                                                                          int(x[6:8]),\n",
    "                                                                          int(x[8:10]),\n",
    "                                                                          int(x[10:12]),\n",
    "                                                                          int(x[12:])))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.2 提取特征：年，月，日"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# create tfa_year, tfa_month, tfa_day feature\n",
    "df['tfa_year'] = np.array([x.year for x in tfa])\n",
    "df['tfa_month'] = np.array([x.month for x in tfa])\n",
    "df['tfa_day'] = np.array([x.day for x in tfa])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.3 提取特征：weekday\n",
    "- 对结果进行one hot encoding编码"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#isoweekday() 可以返回一周的星期几，e.g.星期日：0；星期一：1\n",
    "df['tfa_wd'] = np.array([x.isoweekday() for x in tfa]) \n",
    "df_tfa_wd = pd.get_dummies(df.tfa_wd, prefix = 'tfa_wd')  # one hot encoding \n",
    "df = pd.concat((df, df_tfa_wd), axis = 1) #添加df['tfa_wd'] 编码后的特征\n",
    "df.drop(['tfa_wd'], axis = 1, inplace = True)#删除原有未编码的特征"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.4 提取特征：季节\n",
    "- 因为判断季节关注的是月份，故对年份进行统一"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Y = 2000\n",
    "seasons = [(0, (date(Y,  1,  1),  date(Y,  3, 20))),  #'winter'\n",
    "           (1, (date(Y,  3, 21),  date(Y,  6, 20))),  #'spring'\n",
    "           (2, (date(Y,  6, 21),  date(Y,  9, 22))),  #'summer'\n",
    "           (3, (date(Y,  9, 23),  date(Y, 12, 20))),  #'autumn'\n",
    "           (0, (date(Y, 12, 21),  date(Y, 12, 31)))]  #'winter'\n",
    "\n",
    "def get_season(dt):\n",
    "    dt = dt.date() #获取日期\n",
    "    dt = dt.replace(year=Y) #将年统一换成2000年\n",
    "    return next(season for season, (start, end) in seasons if start <= dt <= end)\n",
    "\n",
    "df['tfa_season'] = np.array([get_season(x) for x in tfa])\n",
    "df_tfa_season = pd.get_dummies(df.tfa_season, prefix = 'tfa_season') # one hot encoding \n",
    "df = pd.concat((df, df_tfa_season), axis = 1)\n",
    "df.drop(['tfa_season'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. date_account_created\n",
    "2.1 将date_account_created转换为datetime类型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dac = pd.to_datetime(df.date_account_created)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2.2 提取特征：年，月，日"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# create year, month, day feature for dac\n",
    "\n",
    "df['dac_year'] = np.array([x.year for x in dac])\n",
    "df['dac_month'] = np.array([x.month for x in dac])\n",
    "df['dac_day'] = np.array([x.day for x in dac])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2.3 提取特征：weekday"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# create features of weekday for dac\n",
    "\n",
    "df['dac_wd'] = np.array([x.isoweekday() for x in dac])\n",
    "df_dac_wd = pd.get_dummies(df.dac_wd, prefix = 'dac_wd')\n",
    "df = pd.concat((df, df_dac_wd), axis = 1)\n",
    "df.drop(['dac_wd'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2.4 提取特征：季节"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# create season features fro dac\n",
    "\n",
    "df['dac_season'] = np.array([get_season(x) for x in dac])\n",
    "df_dac_season = pd.get_dummies(df.dac_season, prefix = 'dac_season')\n",
    "df = pd.concat((df, df_dac_season), axis = 1)\n",
    "df.drop(['dac_season'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2.5提取特征：date_account_created和timestamp_first_active之间的差值\n",
    "- 即用户在airbnb平台活跃到正式注册所花的时间"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dt_span = dac.subtract(tfa).dt.days "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-1     275369\n",
       " 0          7\n",
       " 6          4\n",
       " 5          4\n",
       " 1          4\n",
       " 2          3\n",
       " 3          3\n",
       " 4          3\n",
       " 28         3\n",
       " 94         2\n",
       "dtype: int64"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt_span.value_counts().head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "分析：数据主要集中在-1，可以猜测，用户当天注册dt_span值便是-1\n",
    "\n",
    "    从差值提取特征：差值为一天，一月，一年和其他\n",
    "    即用户活跃到注册花费的时间为一天，一月，一年或其他"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# create categorical feature: span = -1; -1 < span < 30; 31 < span < 365; span > 365\n",
    "def get_span(dt):\n",
    "    # dt is an integer\n",
    "    if dt == -1:\n",
    "        return 'OneDay'\n",
    "    elif (dt < 30) & (dt > -1):\n",
    "        return 'OneMonth'\n",
    "    elif (dt >= 30) & (dt <= 365):\n",
    "        return 'OneYear'\n",
    "    else:\n",
    "        return 'other'\n",
    "\n",
    "df['dt_span'] = np.array([get_span(x) for x in dt_span])\n",
    "df_dt_span = pd.get_dummies(df.dt_span, prefix = 'dt_span')\n",
    "df = pd.concat((df, df_dt_span), axis = 1)\n",
    "df.drop(['dt_span'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#对timestamp_first_active，date_account_created进行特征提取后，从特征列表中删除原有的特征\n",
    "df.drop(['date_account_created','timestamp_first_active'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## age"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Age 获取年龄\n",
    "av = df.age.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在数据探索阶段，我们发现大部分数据是集中在（15，90）区间的，但有部分年龄分布在（1900，2000）区间，我们猜测用户是把出生日期误填为年龄，故进行预处理"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in less\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in greater\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "#This are birthdays instead of age (estimating age by doing 2014 - value)\n",
    "#数据来自2014年，故用2014-value\n",
    "av = np.where(np.logical_and(av<2000, av>1900), 2014-av, av) \n",
    "df['age'] = av"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3.1 将年龄进行分段"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Age has many abnormal values that we need to deal with. \n",
    "age = df.age\n",
    "age.fillna(-1, inplace = True) #空值填充为-1\n",
    "div = 15\n",
    "def get_age(age):\n",
    "    # age is a float number  将连续型转换为离散型\n",
    "    if age < 0:\n",
    "        return 'NA' #表示是空值\n",
    "    elif (age < div):\n",
    "        return div #如果年龄小于15岁，那么返回15岁\n",
    "    elif (age <= div * 2):\n",
    "        return div*2 #如果年龄大于15小于等于30岁，则返回30岁\n",
    "    elif (age <= div * 3):\n",
    "        return div * 3\n",
    "    elif (age <= div * 4):\n",
    "        return div * 4\n",
    "    elif (age <= div * 5):\n",
    "        return div * 5\n",
    "    elif (age <= 110):\n",
    "        return div * 6\n",
    "    else:\n",
    "        return 'Unphysical' #非正常年龄"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#将分段后的年龄作为新的特征放入特征列表中\n",
    "df['age'] = np.array([get_age(x) for x in age])\n",
    "df_age = pd.get_dummies(df.age, prefix = 'age')\n",
    "df = pd.concat((df, df_age), axis = 1)\n",
    "df.drop(['age'], axis = 1, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. 其他特征\n",
    "在数据探索时，我们发现剩余的特征lables都比较少，故不进一步进行特征提取，只进行one-hot-encoding处理"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "feat_toOHE = ['gender', \n",
    "             'signup_method', \n",
    "             'signup_flow', \n",
    "             'language', \n",
    "             'affiliate_channel', \n",
    "             'affiliate_provider', \n",
    "             'first_affiliate_tracked', \n",
    "             'signup_app', \n",
    "             'first_device_type', \n",
    "             'first_browser']\n",
    "#对其他特征进行one-hot-encoding处理\n",
    "for f in feat_toOHE:\n",
    "    df_ohe = pd.get_dummies(df[f], prefix=f, dummy_na=True)\n",
    "    df.drop([f], axis = 1, inplace = True)\n",
    "    df = pd.concat((df, df_ohe), axis = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4.3 整合提取的所有特征\n",
    "\n",
    "    我们将对session以及train，test文件中提取的特征进行合并\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#将对session提取的特征整合到一起\n",
    "df_all = pd.merge(df, df_agg_sess, how='left')\n",
    "df_all = df_all.drop(['id'], axis=1) #删除id\n",
    "df_all = df_all.fillna(-2)  #对没有sesssion data的特征进行缺失值处理\n",
    "\n",
    "#加了一列，表示每一行总共有多少空值，这也作为一个特征\n",
    "df_all['all_null'] = np.array([sum(r<0) for r in df_all.values]) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5. 模型构建"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "5.1 数据准备\n",
    "1. 将train和test数据进行分离操作\n",
    "2. train_row是之前记录的train数据行数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Xtrain = df_all.iloc[:train_row, :]\n",
    "Xtest = df_all.iloc[train_row:, :]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2. 将提取的特征生成csv文件"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Xtrain.to_csv(\"Airbnb_xtrain_v2.csv\")\n",
    "Xtest.to_csv(\"Airbnb_xtest_v2.csv\")\n",
    "#labels.tofile（）：Write array to a file as text or binary (default)\n",
    "labels.tofile(\"Airbnb_ytrain_v2.csv\", sep='\\n', format='%s') #存放目标变量"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "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>tfa_year</th>\n",
       "      <th>tfa_month</th>\n",
       "      <th>tfa_day</th>\n",
       "      <th>tfa_wd_1</th>\n",
       "      <th>tfa_wd_2</th>\n",
       "      <th>tfa_wd_3</th>\n",
       "      <th>tfa_wd_4</th>\n",
       "      <th>tfa_wd_5</th>\n",
       "      <th>tfa_wd_6</th>\n",
       "      <th>tfa_wd_7</th>\n",
       "      <th>...</th>\n",
       "      <th>c_284</th>\n",
       "      <th>c_285</th>\n",
       "      <th>c_286</th>\n",
       "      <th>c_287</th>\n",
       "      <th>c_288</th>\n",
       "      <th>c_289</th>\n",
       "      <th>c_290</th>\n",
       "      <th>c_291</th>\n",
       "      <th>c_292</th>\n",
       "      <th>all_null</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2009</td>\n",
       "      <td>3</td>\n",
       "      <td>19</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>...</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2009</td>\n",
       "      <td>5</td>\n",
       "      <td>23</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>...</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2009</td>\n",
       "      <td>6</td>\n",
       "      <td>9</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>...</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2009</td>\n",
       "      <td>10</td>\n",
       "      <td>31</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>...</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2009</td>\n",
       "      <td>12</td>\n",
       "      <td>8</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>...</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>293</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 497 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   tfa_year  tfa_month  tfa_day  tfa_wd_1  tfa_wd_2  tfa_wd_3  tfa_wd_4  \\\n",
       "0      2009          3       19         0         0         0         1   \n",
       "1      2009          5       23         0         0         0         0   \n",
       "2      2009          6        9         0         1         0         0   \n",
       "3      2009         10       31         0         0         0         0   \n",
       "4      2009         12        8         0         1         0         0   \n",
       "\n",
       "   tfa_wd_5  tfa_wd_6  tfa_wd_7    ...     c_284  c_285  c_286  c_287  c_288  \\\n",
       "0         0         0         0    ...      -2.0   -2.0   -2.0   -2.0   -2.0   \n",
       "1         0         1         0    ...      -2.0   -2.0   -2.0   -2.0   -2.0   \n",
       "2         0         0         0    ...      -2.0   -2.0   -2.0   -2.0   -2.0   \n",
       "3         0         1         0    ...      -2.0   -2.0   -2.0   -2.0   -2.0   \n",
       "4         0         0         0    ...      -2.0   -2.0   -2.0   -2.0   -2.0   \n",
       "\n",
       "   c_289  c_290  c_291  c_292  all_null  \n",
       "0   -2.0   -2.0   -2.0   -2.0       293  \n",
       "1   -2.0   -2.0   -2.0   -2.0       293  \n",
       "2   -2.0   -2.0   -2.0   -2.0       293  \n",
       "3   -2.0   -2.0   -2.0   -2.0       293  \n",
       "4   -2.0   -2.0   -2.0   -2.0       293  \n",
       "\n",
       "[5 rows x 497 columns]"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#读取特征文件\n",
    "xtrain = pd.read_csv(\"Airbnb_xtrain_v2.csv\",index_col=0)\n",
    "ytrain = pd.read_csv(\"Airbnb_ytrain_v2.csv\", header=None)\n",
    "xtrain.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NDF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NDF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>US</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>US</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       0\n",
       "0    NDF\n",
       "1    NDF\n",
       "2     US\n",
       "3  other\n",
       "4     US"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ytrain.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3. 将目标变量进行labels encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\label.py:111: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    }
   ],
   "source": [
    "le = LabelEncoder()\n",
    "ytrain_le = le.fit_transform(ytrain.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 7,  7, 10, ...,  7,  7,  7], dtype=int64)"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ytrain_le"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "4. 提取10%的数据进行模型训练\n",
    "\n",
    "    减少训练模型花费的时间"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Let us take 10% of the data for faster training. \n",
    "n = int(xtrain.shape[0]*0.1)\n",
    "xtrain_new = xtrain.iloc[:n, :]  #训练数据\n",
    "ytrain_new = ytrain_le[:n]       #训练数据的目标变量"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. StandardScaling the dataset\n",
    "- Standardization of a dataset is a common requirement for many machine learning estimators: they might behave badly if the individual feature do not more or less look like standard normally distributed data (e.g. Gaussian with 0 mean and unit variance)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_scaler = StandardScaler()\n",
    "xtrain_new = X_scaler.fit_transform(xtrain_new)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.2 评分模型：NDCG\n",
    "\n",
    "- NDCG是一种衡量排序质量的评价指标，该指标考虑了所有元素的相关性\n",
    "- 由于我们预测的目标变量并不是二分类变量，故我们用NDGG模型来进行模型评分，判断模型优劣\n",
    "- 一般二分类变量: 我们习惯于使用 f1 score, precision, recall, auc score来进行模型评分\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import make_scorer\n",
    "\n",
    "def dcg_score(y_true, y_score, k=5):\n",
    "    \n",
    "    \"\"\"\n",
    "    y_true : array, shape = [n_samples] #数据\n",
    "        Ground truth (true relevance labels).\n",
    "    y_score : array, shape = [n_samples, n_classes] #预测的分数\n",
    "        Predicted scores.\n",
    "    k : int\n",
    "    \"\"\"\n",
    "    order = np.argsort(y_score)[::-1] #分数从高到低排序\n",
    "    y_true = np.take(y_true, order[:k]) #取出前k[0,k）个分数\n",
    "      \n",
    "    gain = 2 ** y_true - 1   \n",
    "\n",
    "    discounts = np.log2(np.arange(len(y_true)) + 2)\n",
    "    return np.sum(gain / discounts)\n",
    "  \n",
    "\n",
    "def ndcg_score(ground_truth, predictions, k=5):   \n",
    "\n",
    "    \"\"\"\n",
    "    Parameters\n",
    "    ----------\n",
    "    ground_truth : array, shape = [n_samples]\n",
    "        Ground truth (true labels represended as integers).\n",
    "    predictions : array, shape = [n_samples, n_classes] \n",
    "        Predicted probabilities. 预测的概率\n",
    "    k : int\n",
    "        Rank.\n",
    "    \"\"\"\n",
    "    lb = LabelBinarizer()\n",
    "    lb.fit(range(len(predictions) + 1))\n",
    "    T = lb.transform(ground_truth)    \n",
    "    scores = []\n",
    "    # Iterate over each y_true and compute the DCG score\n",
    "    for y_true, y_score in zip(T, predictions):\n",
    "        actual = dcg_score(y_true, y_score, k)\n",
    "        best = dcg_score(y_true, y_true, k)\n",
    "        score = float(actual) / float(best)\n",
    "        scores.append(score)\n",
    "\n",
    "    return np.mean(scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. 构建模型\n",
    "### 6.1 Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The training score is:0.7595157690333219\n",
      "\n",
      "The cv score is :0.7417455860527811\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(C = 1.0, penalty='l2', multi_class='ovr')\n",
    "RANDOM_STATE = 2019 #随机种子\n",
    "\n",
    "#k-fold cross validation(k-折叠交叉验证)\n",
    "kf=KFold(n_splits=5,random_state=RANDOM_STATE) # 分成5个组\n",
    "train_score=[]\n",
    "cv_score=[]\n",
    "\n",
    "#select a k \n",
    "k_ndcg = 3\n",
    "#kf.split:Generate indices to split data into training and test set.\n",
    "for train_index,test_index in kf.split(xtrain_new,ytrain_new):\n",
    "    #训练集数据分割为训练集和测试集，y是目标变量\n",
    "    X_train,X_test=xtrain_new[train_index,:],xtrain_new[test_index,:]\n",
    "    y_train,y_test=ytrain_new[train_index],ytrain_new[test_index]\n",
    "    \n",
    "    lr.fit(X_train,y_train)\n",
    "    \n",
    "    y_pred = lr.predict_proba(X_test)\n",
    "    train_ndcg_score = ndcg_score(y_train,lr.predict_proba(X_train),k=k_ndcg)\n",
    "    cv_ndcg_score = ndcg_score(y_test,y_pred,k=k_ndcg)\n",
    "    \n",
    "    train_score.append(train_ndcg_score)\n",
    "    cv_score.append(cv_ndcg_score)\n",
    "\n",
    "print(\"\\nThe training score is:{}\".format(np.mean(train_score)))\n",
    "print(\"\\nThe cv score is :{}\".format(np.mean(cv_score)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "learning curve of logistic regression\n",
    "\n",
    "    观察逻辑回归模型学习曲线的变化\n",
    "    1. 改变逻辑回归参数iteration\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\utils\\optimize.py:203: ConvergenceWarning: newton-cg failed to converge. Increase the number of iterations.\n",
      "  \"number of iterations.\", ConvergenceWarning)\n"
     ]
    }
   ],
   "source": [
    "#set the iterations\n",
    "iteration =[1,5,10,15,20,50,100]\n",
    "\n",
    "kf=KFold(n_splits=3,random_state=RANDOM_STATE)\n",
    "\n",
    "train_score=[]\n",
    "cv_score=[]\n",
    "\n",
    "#select a k:\n",
    "k_ndcg=5\n",
    "\n",
    "for i,item in enumerate(iteration):\n",
    "    lr = LogisticRegression(C=1.0,max_iter=item,tol=1e-5,solver='newton-cg',multi_class='ovr')\n",
    "    train_score_iter = []\n",
    "    cv_score_iter = []\n",
    "    \n",
    "    for train_index,test_index in kf.split(xtrain_new,ytrain_new):\n",
    "        X_train,X_test = xtrain_new[train_index,:],xtrain_new[test_index,:]\n",
    "        y_train,y_test = ytrain_new[train_index],ytrain_new[test_index]\n",
    "        \n",
    "        lr.fit(X_train,y_train)\n",
    "        \n",
    "        y_pred = lr.predict_proba(X_test)\n",
    "        train_ndcg_score = ndcg_score(y_train,lr.predict_proba(X_train),k = k_ndcg)\n",
    "        cv_ndcg_score = ndcg_score(y_test,y_pred,k=k_ndcg)\n",
    "        \n",
    "        train_score_iter.append(train_ndcg_score)\n",
    "        cv_score_iter.append(cv_ndcg_score)\n",
    "    \n",
    "    train_score.append(np.mean(train_score_iter))\n",
    "    cv_score.append(np.mean(cv_score_iter))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAoAAAAEYCAYAAADMEEeQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSJZRAigiIiKSZf4/KbN8GF/ZF0YAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c13feeef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ymin=np.min(cv_score)-0.05\n",
    "ymax = np.max(train_score)+0.05\n",
    "\n",
    "plt.figure(figsize=(9,4))\n",
    "plt.plot(iteration,train_score,'ro-',label='training')\n",
    "plt.plot(iteration,cv_score,'b*-',label='Cross-validation')\n",
    "plt.xlabel('iterations')\n",
    "plt.ylabel('Score')\n",
    "plt.xlim(-5,np.max(iteration)+10)\n",
    "plt.ylim(ymin,ymax)\n",
    "plt.plot(np.linspace(20,20,50),np.linspace(ymin,ymax,50),'g--')\n",
    "plt.legend(loc='lower right',fontsize=12)\n",
    "plt.title(\"Score vs iteration learning curve\")\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "分析：随着iteration的增大，逻辑回归模型的评分在不断升高，当iteration超过20的时候，模型的评分基本不变"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2. 改变数据量大小"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(142, 497) (71, 497)\n",
      "(142, 497) (71, 497)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\utils\\optimize.py:203: ConvergenceWarning: newton-cg failed to converge. Increase the number of iterations.\n",
      "  \"number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(142, 497) (71, 497)\n",
      "(284, 497) (142, 497)\n",
      "(284, 497) (142, 497)\n",
      "(284, 497) (142, 497)\n",
      "(711, 497) (356, 497)\n",
      "(711, 497) (356, 497)\n",
      "(712, 497) (355, 497)\n",
      "(1422, 497) (712, 497)\n",
      "(1423, 497) (711, 497)\n",
      "(1423, 497) (711, 497)\n",
      "(2846, 497) (1423, 497)\n",
      "(2846, 497) (1423, 497)\n",
      "(2846, 497) (1423, 497)\n",
      "(7114, 497) (3558, 497)\n",
      "(7115, 497) (3557, 497)\n",
      "(7115, 497) (3557, 497)\n",
      "(14230, 497) (7115, 497)\n",
      "(14230, 497) (7115, 497)\n",
      "(14230, 497) (7115, 497)\n"
     ]
    }
   ],
   "source": [
    "# Chaning the sampling size\n",
    "# set the iter to the best iteration: iter = 20\n",
    "\n",
    "perc = [0.01,0.02,0.05,0.1,0.2,0.5,1]\n",
    "\n",
    "kf = KFold(n_splits=3, random_state=RANDOM_STATE)\n",
    "\n",
    "train_score = []\n",
    "cv_score = []\n",
    "\n",
    "# select a k:\n",
    "k_ndcg = 5\n",
    "\n",
    "for i, item in enumerate(perc):\n",
    "    \n",
    "    lr = LogisticRegression(C=1.0, max_iter=20, tol=1e-6, solver='newton-cg', multi_class='ovr')\n",
    "    train_score_iter = []\n",
    "    cv_score_iter = []\n",
    "    \n",
    "    n = int(xtrain_new.shape[0]*item)\n",
    "    xtrain_perc = xtrain_new[:n, :]\n",
    "    ytrain_perc = ytrain_new[:n]\n",
    "\n",
    "\n",
    "    for train_index, test_index in kf.split(xtrain_perc, ytrain_perc):\n",
    "        \n",
    "        X_train, X_test = xtrain_perc[train_index, :], xtrain_perc[test_index, :]\n",
    "        y_train, y_test = ytrain_perc[train_index], ytrain_perc[test_index]\n",
    "\n",
    "        print(X_train.shape, X_test.shape)\n",
    "        \n",
    "        lr.fit(X_train, y_train)\n",
    "\n",
    "        y_pred = lr.predict_proba(X_test)\n",
    "        train_ndcg_score = ndcg_score(y_train, lr.predict_proba(X_train), k = k_ndcg)\n",
    "        cv_ndcg_score = ndcg_score(y_test, y_pred, k=k_ndcg)\n",
    "\n",
    "        train_score_iter.append(train_ndcg_score)\n",
    "        cv_score_iter.append(cv_ndcg_score)\n",
    "        \n",
    "    train_score.append(np.mean(train_score_iter))\n",
    "    cv_score.append(np.mean(cv_score_iter))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAoAAAAEYCAYAAADMEEeQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrrLQHTQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c13fee390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ymin = np.min(cv_score)-0.1\n",
    "ymax = np.max(train_score)+0.1\n",
    "\n",
    "plt.figure(figsize=(9,4))\n",
    "plt.plot(np.array(perc)*100, train_score, 'ro-', label = 'training')\n",
    "plt.plot(np.array(perc)*100, cv_score, 'bo-', label = 'Cross-validation')\n",
    "plt.xlabel(\"Sample size (unit %)\")\n",
    "plt.ylabel(\"Score\")\n",
    "plt.xlim(-5, np.max(perc)*100+10)\n",
    "plt.ylim(ymin, ymax)\n",
    "\n",
    "plt.legend(loc = 'lower right', fontsize = 12)\n",
    "plt.title(\"Score vs sample size learning curve\")\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "分析：随着数据量的增加，逻辑回归模型对测试集的预测评分（蓝色线）在不断上升，因为我们在训练模型时只用了10%的数据，如果使用全部的数据，效果可能会更好"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "6.2 树模型\n",
    "\n",
    "    其中的模型包括DecisionTree，RandomForest，AdaBoost，Bagging，ExtraTree，GraBoost\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import AdaBoostClassifier,BaggingClassifier,ExtraTreesClassifier\n",
    "from sklearn.ensemble import GradientBoostingClassifier,RandomForestClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import *\n",
    "from sklearn.svm import SVC ,LinearSVC,NuSVC\n",
    "LEARNING_RATE = 0.1\n",
    "N_ESTIMATORS = 50\n",
    "RANDOM_STATE = 2019\n",
    "MAX_DEPTH = 9\n",
    "\n",
    "#建了一个tree字典\n",
    "clf_tree={\n",
    "    'DTree':DecisionTreeClassifier(max_depth=MAX_DEPTH,random_state=RANDOM_STATE),\n",
    "    'RF':RandomForestClassifier(n_estimators=N_ESTIMATORS,max_depth=MAX_DEPTH,random_state=RANDOM_STATE),\n",
    "    'AdaBoost':BaggingClassifier(n_estimators=N_ESTIMATORS,random_state=RANDOM_STATE),\n",
    "    'ExtraTree':ExtraTreesClassifier(max_depth=MAX_DEPTH,n_estimators=N_ESTIMATORS,random_state=RANDOM_STATE),\n",
    "    'GraBoost':GradientBoostingClassifier(learning_rate=LEARNING_RATE,max_depth=MAX_DEPTH,n_estimators=N_ESTIMATORS,random_state=RANDOM_STATE)\n",
    "}\n",
    "train_score=[]\n",
    "cv_score=[]\n",
    "\n",
    "kf=KFold(n_splits=3,random_state=RANDOM_STATE)\n",
    "\n",
    "k_ndcg=5\n",
    "\n",
    "for key in clf_tree.keys():\n",
    "    clf=clf_tree.get(key)\n",
    "    \n",
    "    train_score_iter=[]\n",
    "    cv_score_iter=[]\n",
    "    \n",
    "    for train_index,test_index in kf.split(xtrain_new,ytrain_new):\n",
    "        X_train,X_test = xtrain_new[train_index,:],xtrain_new[test_index,:]\n",
    "        y_train,y_test = ytrain_new[train_index],ytrain_new[test_index]\n",
    "        \n",
    "        clf.fit(X_train,y_train)\n",
    "        \n",
    "        y_pred = clf.predict_proba(X_test)\n",
    "        train_ndcg_score = ndcg_score(y_train,clf.predict_proba(X_train),k=k_ndcg)\n",
    "        cv_ndcg_score = ndcg_score(y_test,y_pred,k=k_ndcg)\n",
    "        \n",
    "        train_score_iter.append(train_ndcg_score)\n",
    "        cv_score_iter.append(cv_ndcg_score)\n",
    "    \n",
    "    train_score.append(np.mean(train_score_iter))\n",
    "    cv_score.append(np.mean(cv_score_iter))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAFgCAYAAAC2QAPxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzs3Xd8VFX6x/HPE0oooUmHQEBXEVRs\nseCubS2IiiBWimUt/CyorA0FF5FiQQELuC4qtl+UtWP9qazLrqwVV0EBQURCiSi9RQMh5/fHmcCk\nFyZzp3zfr1deM3PvnTvPDMPMM+c85xxzziEiIiKSSFKCDkBEREQk0pTgiIiISMJRgiMiIiIJRwmO\niIiIJBwlOCIiIpJwlOCIiIhIwlGCI5KgzOwxM/tL2O2rzexnM9tqZs3N7Pdm9n3odt8gY00GZjbK\nzP43Que61MxmR+JcIolKCY5IHDKzZWb2q5ltMbONZvaxmV1lZrv+TzvnrnLOjQkdXweYCJzqnEtz\nzq0DRgOTQ7dfj3L8FX5Bm9ksM7siWjFFkpmdYGYrg45DJJkpwRGJX72dc42ADOBeYBjwZBnHtgbq\nAfPDtmUUu11pZla7OveLpFiIQURilxIckTjnnNvknHsDuAC4xMwOBDCzp81srJntBywKHb7RzD40\nsx+AvYE3Q11UqWbWxMyeNLOfzGxV6L61Que61Mz+Y2aTzGw9MCq0/TIzW2hmG8zsPTPLKIzLzFyo\nVen70P4p5nUFHgN6hB57Y/HnZGbjgGOByaFjJoed81oz+x74PrRtfzP7wMzWm9kiMzs/7DypZvaA\nmS0Pdc89Zmb1S3sdiz3HjWa21MyOCW1fYWa/mNklFZ3bzBoC7wLtQrFvNbN2obvVNbNnQy1v880s\nM+x8XUOtVhtD+84K29fczN4ws81m9jmwT9g+C8X8i5ltMrN5he8BkWSmBEckQTjnPgdW4hOD8O2L\ngQNCN5s65/7onNsHWI5vBUpzzuUBzwD5wO+AQ4FTgfAuoqOApUArYFyobmc40A9oCXwEvFAsrDOB\nI4CDgfOBns65hcBVwCehx25aynMZETrfkNAxQ8J29w3F0i2UTHwAPB+Kqz/wqJkVPt/7gP2AQ0LP\nqz0wsuxXkaOAeUDz0Dmnh+L/HTAIn3CllXdu59w2oBeQE4o9zTmXE7rPWaFzNgXeAAoTtzrAm8D7\noedxHZBlZl1C95sC/Aa0BS4L/RU6FTguFEtTfKK7rpznKJIUlOCIJJYcYK+q3snMWuO/lIc657Y5\n534BJgEXhp/bOfeIcy7fOfcr8D/APc65hc65fOBu4JDwVhzgXufcRufccuCf+GRgT93jnFsfiuFM\nYJlz7qlQXP8FXgHONTMDrgT+HDp+SyjGC8s+NT+GzrUT+DvQARjtnMtzzr0PbAd+V81zA8x2zr0T\nOv9z+MQP4GggDf96bXfOfQi8BfQPtaKdQyh5cs59i09GC+0AGgH7Axb69/ipwldRJMGpD1sksbQH\n1lfjfhlAHeAn/90N+B9AK8KOWVHKfR4yswlh2ywUQ3bo9uqwfbn4L/E9FR5HBnBUsW6u2vjkoSXQ\nAPgy7DkZUKucc/8cdv1XAOdc8W1p1Tw3lHw96oVqidoBK5xzBWH7s/GvZcvQc1pRbB+h+D4MdeFN\nATqa2WvAzc65zRXEIpLQlOCIJAgzOwL/hVid4cMrgDygRag1pjSulPuMc85lVePxip+rKseEb18B\n/Ms5d0rxg0Ijyn4FDnDOrap6iOVaW8G5K/P8wuUAHcwsJSzJ6QgsBtbguw47AN+F7dv9YM49DDxs\nZq2AF4FbgL8gksTURSUS58yssZmdia/t+F/n3DdVPUeoS+N9YELofClmto+ZHV/O3R4Dbi+sdwkV\nKZ9XyYf8GUg3s7oVHLN3Bed5C9jPzC4yszqhvyPMrGsoUXgcmBT64sfM2ptZz0rGWKZKnPtnoLmZ\nNankKT8DtgG3hp7DCUBvYHqoO+tVYJSZNTCzbkB4sfMRZnZUqI5nG75WZ+eePkeReKcERyR+vWlm\nW/CtGCPw89z8aQ/OdzFQF1gAbABexhe1lso59xq+0Ha6mW0GvsXX8VTGh/gh6qvNbG0ZxzyEr6XZ\nYGYPlxHDFnyR7YX4VpDVoZhSQ4cMA5YAn4ZinAl0KeVU1VHmuZ1z3+ELrpeGRkW1K/s04Jzbji9A\n7oVvHXoUuDh0HoAh+K6x1cDTwFNhd2+MT7Y24Luu1gEPROD5icQ1c66qLakiIiIisU0tOCIiIpJw\nlOCIiIhIwlGCIyIiIglHCY6IiIgknIScB6dFixauU6dOQYchIiIiEfbll1+udc61rOi4hExwOnXq\nxJw5c4IOQ0RERCLMzLIrPkpdVCIiIpKAlOCIiIhIwlGCIyIiIglHCY6IiIgknIQsMhYRkdhTUFDA\nypUr2bZtW9ChSIxr2LAh6enppKRUvx1GCY6IiETF2rVrMTO6dOmyR19cktgKCgpYtWoVa9eupVWr\nVtU+j95hIiISFRs3bqR169ZKbqRcKSkptG7dmk2bNu3ZeSIUj4iISLl27txJnTp1gg5D4kCdOnXI\nz8/fo3MowRERkagxs6BDkDgQifeJEhwRERFJOEpwRERE9tBVV13FmDFjIn6sVJ9GUYmISNLr1KkT\nTzzxBCeffHK17v/YY4/VyLFSfWrBERGR2JaVBZ06QUqKv8zKiurD72mxqwRDCY6IiMSurCwYPBiy\ns8E5fzl4cESTnIsuuojly5fTu3dv0tLSGD9+PGbGk08+SceOHfnjH/8IwHnnnUebNm1o0qQJxx13\nHPPnz991jksvvZQ77rgDgFmzZpGens6ECRNo1aoVbdu25amnnqrWsevWraN37940btyYI444gjvu\nuIM//OEPEXvuiUxdVCIiEn1Dh8LXX1d83KefQl5e0W25uXD55fD44+Xf95BD4MEHK3yI5557jo8+\n+mhXF9WyZcsYNmwY//rXv1i4cOGueXt69erFtGnTqFu3LsOGDWPgwIF8XcZzWL16NZs2bWLVqlV8\n8MEHnHvuufTt25dmzZpV6dhrr72Whg0bsnr1apYtW0bPnj3JyMio8DmJWnBERCSWFU9uKtoeQaNG\njaJhw4bUr18fgMsuu4xGjRqRmprKqFGjmDt3bpmT0dWpU4eRI0dSp04dTj/9dNLS0li0aFGVjt25\ncyevvPIKd911Fw0aNKBbt25ccsklNfZ8E41acEREJPoq0bIC+Jqb7OyS2zMyYNasSEZUQocOHXZd\n37lzJyNGjOCll15izZo1u1p11q5dS5MmTUrct3nz5tSuvfsrtkGDBmzdurXUxynr2DVr1pCfn18k\njvDrUj614IiISOwaNw4aNCi6rUEDvz2CSptYLnzb888/z4wZM5g5cyabNm1i2bJlADjnIhpHuJYt\nW1K7dm1Wrly5a9uKFStq7PESjRIcERGJXQMHwtSpvsXGzF9Oneq3R1Dr1q1ZunRpmfu3bNlCamoq\nzZs3Jzc3l+HDh0f08UtTq1Yt+vXrx6hRo8jNzeW7777j2WefrfHHTRRKcEREJLYNHAjLlkFBgb+M\ncHIDcPvttzN27FiaNm3Kyy+/XGL/xRdfTEZGBu3bt6dbt24cffTREY+hNJMnT2bTpk20adOGiy66\niP79+5OamhqVx453VpPNa0HJzMx0c+bMCToMEREJs3DhQrp27Rp0GHFt2LBhrF69mmeeeSboUGpc\nWe8XM/vSOZdZ0f3VgiMiIhKjvvvuO+bNm4dzjs8//5wnn3ySs88+O+iw4oJGUYmIiMSoLVu20L9/\nf3JycmjVqhU33XQTffr0CTqsuKAER0REJEYdccQRLFmyJOgw4pK6qERERCThKMERERGRhKMER0RE\nRBJOVBIcM5tmZr+Y2bdl7Dcze9jMlpjZPDM7LGzfJWb2fehPi3CIiIhIhaLVgvM0cFo5+3sB+4b+\nBgN/BTCzvYA7gaOAI4E7zazkUqwiIiIiYaKS4Djn/g2sL+eQPsCzzvsUaGpmbYGewAfOufXOuQ3A\nB5SfKImIiCS9UaNGMWjQIACWL19OWloaO3furPDY6jjggAOYVcMLn1ZHrNTgtAfCVxBbGdpW1vYS\nzGywmc0xszlr1qypsUBFRCRxPf/882RmZpKWlkbbtm3p1asXs2fPDjqsPdKxY0e2bt1KrVq19vhc\nl156KXfccUeRbfPnz+eEE07Y43NHWqwkOCWXcQVXzvaSG52b6pzLdM5ltmzZMqLBiYhIcLKyoFMn\nSEnxl1lZNfM4EydOZOjQoQwfPpyff/6Z5cuXc8011zBjxowSx+bn59dMEBIxsZLgrAQ6hN1OB3LK\n2S4iNSla3ygiFcjKgsGDITsbnPOXgwdH/i25adMmRo4cyZQpU+jXrx8NGzakTp069O7dm/vvv59R\no0Zx7rnnMmjQIBo3bszTTz9NXl4eQ4cOpV27drRr146hQ4eSl5cHwNq1aznzzDNp2rQpe+21F8ce\neywFBQUA3HfffbRv355GjRrRpUsX/vGPf5Qa02mnncbkyZOLbDv44IN59dVXAbjhhhvo0KEDjRs3\n5vDDD+ejjz4q9TzLli3DzHYlZT/++CPHH388jRo14pRTTmHt2rVFjj/vvPNo06YNTZo04bjjjmP+\n/PkATJ06laysLMaPH09aWhq9e/cGoFOnTsycOROg3Ndk1qxZpKenM2HCBFq1akXbtm156qmnqvYP\nVQWxMpPxG8AQM5uOLyje5Jz7yczeA+4OKyw+Fbg9qCBFkkLhN0purr9d+I0CNbKKsySnoUPh668r\nPu7TTyH0/bhLbi5cfjk8/nj59z3kEHjwwcrF88knn/Dbb7+Vu87TjBkzeOmll3j22WfJy8tj3Lhx\nfPrpp3z99deYGX369GHs2LGMGTOGCRMmkJ6eTmHJxKeffoqZsWjRIiZPnswXX3xBu3btWLZsWZm1\nMQMGDOBvf/sbQ4YMAWDBggVkZ2dzxhlnAH6W45EjR9KkSRMeeughzjvvPJYtW0a9evXKfa4DBgyg\nR48evP/++3z22WecccYZRZZ/6NWrF9OmTaNu3boMGzaMgQMH8vXXXzN48GA+/vhj0tPTGTt2bKnn\nLu81AVi9ejWbNm1i1apVfPDBB5x77rn07duXZs0iP34oWsPEXwA+AbqY2Uozu9zMrjKzq0KHvAMs\nBZYAjwPXADjn1gNjgC9Cf6ND20SkpowYsTu5KZSb67eLRFnx5Kai7dW1bt06WrRoQe3aZf/u79Gj\nB3379iUlJYX69euTlZXFyJEjadWqFS1btuTOO+/kueeeA6BOnTr89NNPZGdnU6dOHY499ljMjFq1\napGXl8eCBQvYsWMHnTp1Yp999in18c4++2y+/vprsrOzAcjKyqJfv36kpqYCMGjQIJo3b07t2rW5\n6aabyMvLY9GiReU+z+XLl/PFF18wZswYUlNTOe6443a1xBS67LLLaNSoEampqYwaNYq5c+eyadOm\nSr2O5b0mha/LyJEjqVOnDqeffjppaWkVxlxdUWnBcc71r2C/A64tY980YFpNxCUipVi+vGrbRaqh\nsi0rnTr5RsTiMjIgkgN3mjdvztq1a8nPzy8zyenQoUOR2zk5OWRkZITFlEFOjq+iuOWWWxg1ahSn\nnnoqAIMHD+a2227jd7/7HQ8++CCjRo1i/vz59OzZk4kTJ9KuXTvS0tJ2nWvBggV07NiRM844g+nT\npzNs2DCmT5/O1KlTdx0zYcIEnnjiCXJycjAzNm/eXKK7qbicnByaNWtGw4YNi8S9YoUfz7Nz505G\njBjBSy+9xJo1a0hJ8e0ga9eupUmTJhW+juW9JsCuhKxQgwYN2Lp1a4XnrY5YqcERkVhR7EN8l44d\noxuHCDBuHDRoUHRbgwZ+eyT16NGDevXq8frrr5d5jFnRcS/t2rXb1boCvnWkXbt2ADRq1IgJEyaw\ndOlS3nzzTSZOnLir1mbAgAHMnj2b7OxszIxhw4YBsHXr1l1/HUP/3/r3788LL7zAJ598wq+//sqJ\nJ54IwEcffcR9993Hiy++yIYNG9i4cSNNmjTBtxeUrW3btmzYsIFt27YVibvQ888/z4wZM5g5cyab\nNm1i2bJlALvOW/w1KK681yTalOCISFGhD9AiauIbRaQSBg6EqVN9i42Zv5w6NfLlYE2aNGH06NFc\ne+21vP766+Tm5rJjxw7effddbr311lLv079/f8aOHcuaNWtYu3Yto0eP3jWfzFtvvcWSJUtwztG4\ncWNq1apFrVq1WLRoER9++CF5eXnUq1eP+vXrlzt8+/TTTyc7O5uRI0dywQUX7GpR2bJlC7Vr16Zl\ny5bk5+czevRoNm/eXOHzzMjIIDMzkzvvvJPt27cze/Zs3nzzzV37t2zZQmpqKs2bNyc3N5fhw4cX\nuX/r1q1ZunRpmecv7zWJNiU4IrLb6tXw2mtwwAG+xabw19pZZ6nAWAIzcCAsWwYFBf6ypt6KN954\nIxMnTmTs2LG0bNmSDh06MHnyZPr27Vvq8XfccQeZmZl0796dgw46iMMOO2zXHDHff/89J598Mmlp\nafTo0YNrrrmGE044gby8PG677TZatGhBmzZt+OWXX7j77rvLjCk1NZV+/foxc+ZMBgwYsGt7z549\n6dWrF/vttx8ZGRnUq1evRBdaWZ5//nk+++wz9tprL+666y4uvvjiXfsuvvhiMjIyaN++Pd26dePo\no48uct/LL7+cBQsW0LRp01Jfl/Jek2izipqz4lFmZqabM2dO0GGIxJ8BA+CVV+Cbb2C//fy2Pn1g\n5kz47ruyu69EKmHhwoV07do16DAkTpT1fjGzL51zmRXdXy04IuLNnAkvvAC33747uQF46CE/AcnQ\nocHFJiJSRUpwRMSPub32Wvjd7+C224ru69QJRo6EV1+Fd94JJDwRkapSgiMiMH48LF4MU6ZAaZOE\n3Xgj7L8/DBkCv/4a/fhERKpICY5IsluyxI+QOv98CM3ZUULduvDoo/Djj3DPPdGNT0SkGpTgiCQz\n53yrTN26MGlS+ceeeCIMGgT33edbe0SqIREHtkjkReJ9ogRHJJm9/DK89x6MHQuVmYzr/vuhfn1f\nr6MvKqmievXqsW7dOiU5Ui7nHOvWratwTa2KaJi4SLLavBm6doXWreHzz6GcNXiKmDLFt/pMnw4X\nXFCzMUpC2bFjBytXruS3334LOhSJcfXq1SM9PZ06deqU2FfZYeJKcESS1dCh8PDDfrnmI4+s/P12\n7oSjjoKcHD83TuPGNRejiEgxmgdHRMr21VfwyCNw1VVVS24AatWCv/7Vz3o8cmTNxCcisoeU4Igk\nm4ICuPpqaNECypkivlxHHOGTo0cega+/jmx8IiIRoARHJNk8/jh89hlMmABNm1b/POPGQfPmPlkq\nKIhcfCIiEaAERySZ/Pyzn6n4xBP3fMXCZs3ggQd8Dc+0aZGJT0QkQpTgiCSTW26Bbdv8pH2FK4Xv\niYsuguOOg2HDYO3aPT+fiEiEKMERSRazZsFzz8Gtt/plFyLBzCdLmzf7JEdEJEYowRFJBtu3+1qZ\nzp1hxIjInvuAA/xaVdOmwX/+E9lzi4hUkxIckWTwwAN+zprJk/1MxJE2ciR06OCTqB07In9+EZEq\nUoIjkuh+/BHGjIFzzoHTT6+Zx2jY0E8a+M03fui4iEjAlOCIJDLn4Lrr/DIMDz5Ys4/Vpw+ccQbc\neSesXFmzjyUiUgElOCKJ7PXX4e234a67ID29Zh/LzLfi5Of7mhwRkQApwRFJVFu3wvXXw8EH+8to\n2HtvuOMOeOklv0q5iEhAlOCIJKpRo3xX0V//WvmVwiPh5pthv/3g2mtBq0aLSECU4IgkonnzfM3N\nlVdCjx7RfezUVD83zg8/wH33RfexRURClOCIJJrCxTSbNYN77w0mhpNOgv794Z57YMmSYGIQkaSm\nBEck0UybBh9/7Oe+2Wuv4OKYMAHq1oUhQ/xoLhGRKIpagmNmp5nZIjNbYma3lbI/w8z+YWbzzGyW\nmaWH7dtpZl+H/t6IVswicWftWr9kwnHHwcUXBxtL27YwdqwvNn7llWBjEZGkE5UEx8xqAVOAXkA3\noL+ZdSt22APAs8657sBo4J6wfb865w4J/Z0VjZhF4tKtt/p1oSK1mOaeuuYaOOQQGDoUtmwJOhoR\nSSLRasE5EljinFvqnNsOTAf6FDumG/CP0PV/lrJfRMrz0Ufw1FNw001+fahYULs2PPYY5OT4UV0i\nIlESrQSnPbAi7PbK0LZwc4FzQtfPBhqZWfPQ7XpmNsfMPjWzvqU9gJkNDh0zZ82aNZGMXST27djh\nC4szMuAvfwk6mqKOOsqP5nroIT+6S0QkCqKV4JTWVl686vBm4Hgz+wo4HlgF5If2dXTOZQIDgAfN\nbJ8SJ3NuqnMu0zmX2bJlywiGLhIHJk2C+fP9OlANGwYdTUl33+1HdV19tR/lJSJSw6KV4KwEOoTd\nTgdywg9wzuU45/o55w4FRoS2bSrcF7pcCswCDo1CzCLxITvbL8XQpw/07h10NKVr3hzGj/eju555\nJuhoRCQJRCvB+QLY18w6m1ld4EKgyGgoM2thZoXx3A5MC21vZmaphccAvwcWRClukdh3ww3+8uGH\ng42jIpdcAn/4A9xyC6xbF3Q0IpLgopLgOOfygSHAe8BC4EXn3HwzG21mhaOiTgAWmdlioDUwLrS9\nKzDHzObii4/vdc4pwREBeOMNmDHDF/B27Bh0NOVLSfGjuzZuhNtvDzoaEUlw5hJwAq7MzEw3Z86c\noMMQqVnbtkG3btC4Mfz3v1CnTtARVc7NN/tJAD/5BI4+OuhoRCTOmNmXobrccmkmY5F4NWYMLF/u\nF9OMl+QGfGtT+/Zw1VWQn1/h4SIi1aEERyQezZ/vW0H+9Cdf1xJP0tL8kPG5c2HKlKCjEZEEpQRH\nJN4454dbN27sRybFo3794LTT/Jw9OTkVHy8SLisLOnXydV2dOvnbIsUowRGJN88842ctHj8eWrQI\nOprqMfNz9mzf7mdeFqmsrCwYPNhPj+Ccvxw8WEmOlKAERySerFvnh1kfc4zvnopnv/sdDB8O06fD\nBx8EHY3EixEjIDe36LbcXL9dJIwSHJF4cvvtsGGDLyxOSYD/vrfe6hOda6+FvLygo5F4sHx51bZL\n0kqAT0iRJPHxx/D44/DnP0P37kFHExn16vlC4++/h/vvDzoaiQdpaaVvj/V5oCTqlOCIxIP8fF9Y\n3KED3Hln0NFE1qmnwvnnw7hxsHRp0NFILJs0CbZs8avUh2vQwL9/RMIowRGJBw8/7Ffifuihsn/B\nxrOJE/2X1nXX+cJRkeJeeskXpJ9zDkybBhkZvlg9IwOmToWBA4OOUGKMZjIWiXUrVkDXrnDCCfDm\nm/5DPRFNmgQ33givvgpnnx10NBJLPvoITjkFMjN9QXr9+kFHJAHSTMYiiWLoUCgo8MOqEzW5Ad96\n0727Xzx069ago5FYsXAh9Onj57uZMUPJjVSaEhyRWPbOO75F4y9/gc6dg46mZtWu7UeHrVgBo0cH\nHY3EgtWroVcvvxTJu+9C8+ZBRyRxRAmOSKzKzYUhQ3z3VLJMhnfMMXD55b676ttvg45GgrRlC5xx\nBqxdC2+/nfgJvkScEhyRWHX33fDjj/Doo1C3btDRRM+99/plKK65RgXHyWrHDj+ybu5cePFFX3sj\nUkVKcERi0cKFfimGiy/2xcXJpEUL/9w/+gieey7oaCTaCtda+7//812Wp58edEQSp5TgiMQa53zr\nRVpa8k5+96c/QY8ecPPNsH590NFINI0ZA08+6evOrrwy6GgkjinBEYk1WVkwaxbccw+0ahV0NMFI\nSfG/3tet0xpDyeTpp/1ElpdcAnfdFXQ0EueU4IjEkg0bfEHxUUfp1+vBB8P118Pf/gaffx50NFLT\n3nvPv+dPOcVP3JfIUyJIVCjBEYklw4f7USOPPZYYi2nuqbvugrZtfU3Gzp1BRyM15auv4Nxz4YAD\n4OWXk6uoXmqMPkFFYsXnn/vWiuuvh0MOCTqa2NC4sR8y/t//+i4rSTzZ2X44eLNmft6nxo2DjkgS\nhJZqEIkF+flw5JHw88/w3XfQqFHQEcUO56BnT/jsM1i0CNq0CToiiZQNG+D3v4ecHPjPf3wLjkgF\ntFSDSDx59FHfTP/gg0puijODyZPht9/8qCpJDHl50Lcv/PADvP66khuJOCU4IkHLyYE77oDTTvN1\nCFLSfvvBbbf5EWYffhh0NLKnCgr8SKl//9uPnEq2uZ4kKpTgiATtz3/2M7dOnqyRI+W57TbYe28/\nR9D27UFHI3ti2DD4+9/9hI79+wcdjSQoJTgiQXr/fT8V/fDhsM8+QUcT2+rX90ngokUwYULQ0Uh1\nPfIIPPAAXHutuhylRqnIWCQov/4KBx0EtWrBvHmQmhp0RPHh3HP9aJv587UAY7x57TU45xw46yx4\n5RX/3hepIhUZi8S6e+/1BZaPPqrkpiomTfJzBN1wQ9CRSFV8/DEMGOAnsXz+eSU3UuOU4IgEYfFi\nn+AMGAAnnRR0NPGlQwcYNQrefBPeeCPoaKQyFi/2rTbp6f7frEGDoCOSJBC1BMfMTjOzRWa2xMxu\nK2V/hpn9w8zmmdksM0sP23eJmX0f+rskWjGL1AjnfP1B/fqqJamuG26AAw/0kyJu2xZ0NFKen3/2\nIwRTUvwK4S1bBh2RJImoJDhmVguYAvQCugH9zaxbscMeAJ51znUHRgP3hO67F3AncBRwJHCnmTWL\nRtwiNeLvf4eZM2HcOE1aV1116viZjbOzYezYoKORsmzbBmeeCatXw1tvqZBeoipaLThHAkucc0ud\nc9uB6UCfYsd0A/4Ruv7PsP09gQ+cc+udcxuAD4DTohCzSORt2uSHhWdmwlVXBR1NfPvDH+DSS/2I\nnAULgo5GisvPhwsv9Mts/P3vfqZukSiKVoLTHlgRdntlaFu4ucA5oetnA43MrHkl74uZDTazOWY2\nZ82aNRELXCSi7rgDfvnFL6apIss9N368n/n52mt915/EhsJu2LfegilToHfvoCOSJBStBKe02cuK\nfxrdDBxvZl8BxwOrgPxK3hfn3FTnXKZzLrOl+nglFn35pR8xdc01cPjhQUeTGFq29MXas2b5kTkS\nG+65B6ZOhdtvV0ulBCZaCc5WbxrPAAAgAElEQVRKoEPY7XQgJ/wA51yOc66fc+5QYERo26bK3Fck\n5u3c6T/oW7VSzUikXXGF7/648UbYuDHoaOS552DECBg40NeZiQQkWgnOF8C+ZtbZzOoCFwJFxnea\nWQszK4zndmBa6Pp7wKlm1ixUXHxqaJtI/HjsMZgzx8/h0qRJ0NEklpQUX3C8dq3vApTgzJwJl10G\nf/wjTJumpUckUFFJcJxz+cAQfGKyEHjROTffzEab2Vmhw04AFpnZYqA1MC503/XAGHyS9AUwOrRN\nJD6sXu2XYjj5ZLjggqCjSUyHHeZrPh591HcFSvTNmwf9+sH++8Orr0LdukFHJElOSzWI1LSBA+Hl\nl+Gbb/yq2FIzNm3yX67p6fDppyrijqYVK6BHD3/9k0/8ZIwiNURLNYjEgn/8wxe/3nabkpua1qQJ\nTJzouwKnTg06muSxcSP06gVbtsC77yq5kZihFhyRmpKXB927+wLjb7+FevWCjijxOee7Ar/80q86\n3rp10BEltrw8n9zMnu2TGy07IlGgFhyRoI0f79fgefRRJTfRYubnXcnNhVtvDTqaxFZQAJdfDv/8\nJzz5pJIbiTmVTnDMLNXMxpnZUjPbFNp2qpkNqbnwROLUDz/4IbLnnw+nnhp0NMll//19cvPss/Cv\nfwUdTeIaMQKysuDuu+Gii4KORqSEqrTgTAIOBAaye6K9+cDVkQ5KJK45B0OG+FEkkyYFHU1yGj4c\nOnXykypu3x50NInnr3/1Eyz+z//4+jKRGFSVBOdsYIBz7hOgAMA5t4pSlk0QSWovv+xXTR47Ftq1\nCzqa5NSgATzyiF+j6sEHg44msbzxhk/gzzwTJk/WXDcSs6qS4GwHaodvMLOWwLqIRiQSzzZvhqFD\n4dBDfeuBBOfMM6FvX7jrLr/quOy5zz7zC2gefjhMnw61a1d8H5GAVCXBeQl4xsw6A5hZW2AyfmVw\nEQG480746Sc/c7E+/INX2HozdGiwcSSCJUv8oplt2/pFNBs2DDoikXJVJcEZDiwDvgGaAt/j14S6\nK/JhicShr76Chx/2dQlHHhl0NAKQkQEjR8Lrr/svZameNWv8cPCCAj8cvFWroCMSqVCl5sEJrRF1\nAvAf51xeqGtqrYvRSXQ0D45EXUEBHHMM/PgjfPcdNGsWdERSaPt232WYmwvz5/v6HKm83Fy/ttTc\nufDhh7tnLBYJSETnwXHOFQAznHN5odtrYjW5EQnE44/7+oQJE5TcxJq6df1cRMuW+SHNUnk7d8KA\nAfD5535GbiU3Ekeq0kX1bzM7usYiEYlXv/zih8qeeKJfd0piz/HH+7laxo/3MxxLxZyD66+HGTN8\n1+vZZwcdkUiVVKUKMht418xmACvYPRcOzrmRkQ5MJG7ccgts2+ZbCTRkNnbdfz+8+aZfdfyDD/Rv\nVZH77/fv6Ztv9sPCReJMVVpw6gOv4xObdKBD6C+9BuISiQ+zZvkZc2+91c+gK7GrdWvfRfWPf8Df\n/x50NLHthRdg2DC44AK4776goxGpFi22KVJd27fDIYfAb7/54tX69YOOSCqycyccfTSsXOmLwZs0\nCTqi2PPPf0LPnr5o/r33IDU16IhEiqiRxTbNbF8zG2lmfwtd7lv9EEXi3IQJsHChn81VyU18qFXL\nLzPw889++LgU9e23vtZm333htdeU3Ehcq8pim72BL4H9gfVAF2COmZ1VQ7GJxK4ff4QxY6BfPzj9\n9KCjkarIzISrr/aJ6VdfBR1N7Fi1yr+XGzSAd97RaECJe5XuojKzb4DrnXP/DNt2AjDZOXdgzYRX\nPeqikhrlnJ/R9V//8i046SpDizsbN0KXLtC5M3z8MaRUqTE78WzeDMceC0uXwkcf+a5XkRhVE11U\n6cBHxbbNRkXGkmxefx3eftuvcaTkJj41beq7GD/7DJ54IuhogrV9O5xzjl+Y9JVXlNxIwqhKgvM1\ncFOxbTeGtoskh61b/dwg3bv7S4lfAwf6+XFuu80vRZCMnIMrr4SZM/1klaeeGnREIhFTlQTnauAK\nM8sxs8/MLAe4MrRdJDncdZcfgfPXv2oxzXhn5ud52bLFD4lORiNH+mkORo+GSy8NOhqRiKp0guOc\n+w7oCpwPTAhddnPOLayh2ERiy7x5MGmS/8V7zDFBRyOR0K2bn8juqadg9uygo4muxx+HsWPhiivg\njjuCjkYk4qpSZHwIsM45tyJsWwdgL+fc3BqKr1pUZCwRV1DgizAXL/ZT/e+1V9ARSaRs2+YTncaN\n4b//hTp1go6o5r39NvTp47ukZsxIjucsCaMmioz/Fyj+v6Au8FxVAhOJS0895Ufb3H+/kptE07Ch\nX2vp22/9ZaKbMwfOPx8OPhhefFHJjSSsqrTgbHbONa7s9iCpBUciau1aP6T4wAP90gxawygxnXUW\nfPihn+E4UUfHLV3qVwRv0AA++QTatAk6IpEqq4kWnJVmdlixBzkMyKlqcCJx5dZb/TwhWkwzsT30\nkO+KHDo06Ehqxrp10KsX7NgB776r5EYSXlUSnEnADDO7zsxON7PrgdeAiTUTmkgMmD3bd0/ddBMc\ncEDQ0UhN6tzZF9u+8opPABLJr7/6FqrsbHjjDS0MK0mhSottmtl5wOX4yf1WAE84516podiqTV1U\nEhE7dsBhh/lhxPPn+1oNSWx5eb42ZccOX5OTCGuM7dzpa25ee83X3Jx7btARieyRiHVRmdnhZnYg\ngHPuJeBiYB7QHjjVzNIqGdBpZrbIzJaY2W2l7O9oZv80s6/MbJ6ZnR7a3snMfjWzr0N/j1Xm8UT2\n2IMP+i+5Rx5RcpMsUlN9V+TSpXDvvUFHs+ecgxtvhFdfhYkTldxIUqlMF9WDQHhn7VTgd8DfgAOA\n8RWdwMxqAVOAXkA3oL+ZdSt22B3Ai865Q4ELgUfD9v3gnDsk9HdVJWIW2TPZ2TBqlB9K27t30NFI\nNP3xjzBggE9wvv8+6Gj2zKRJfmTY0KGJW1skUobKJDhdCa1BZWZNgTOAQc65KUB/oDKf/kcCS5xz\nS51z24HpQJ9ixzigcDRWE1S8LEG64QZ/mQzDhqWkCROgXj0YMsS3gsSjF1/0tWPnnuufj0iSqUyC\nUxvYHrp+NPCTc24xQGjSv6aVOEd7fM1OoZWhbeFGAYPMbCXwDnBd2L7Ooa6rf5nZsZV4PJHqe/NN\nP/nZnXdCx45BRyNBaNMGxo2D99+Hl18OOpqq++gjuOgi+MMf4LnntFq6JKXKvOvnA+eFrl8IzCzc\nYWbtgU2VOEdpY2uL/yzqDzztnEsHTgeeM7MU4CegY6jr6kbgeTMrbT6ewWY2x8zmrEnWhfNkz23b\nBtdd50dM/fnPQUcjQbr6ajj0UN+1s3lz0NFU3sKFvmu1c2efqNerF3REIoGoTIIzDPibma3Hd0/d\nF7bvAuA/lTjHSqBD2O10SnZBXQ68COCc+wSoB7RwzuU559aFtn8J/ADsV/wBnHNTnXOZzrnMli1b\nViIkkVKMGePrbx57TDO8Jrtatfyiqj/95Oux4sFPP/m5burW9UPdNeu2JLEKExzn3GygI3AKsLdz\nblHY7reByvzM/QLY18w6m1ldfEvQG8WOWQ6cBGBmXfEJzhozaxkqUsbM9gb2BZZW4jFFqmb+fF+r\n8Kc/+aZ9kaOOgsGDfS3W3Jhacq+kLVvgjDP8zNtvv+1bcESSWKU6Zp1zW5xzXzrnthTbvsg5V2Ex\nsHMuHxgCvAcsxI+Wmm9mo83srNBhNwFXmtlc4AXgUucn6TkOmBfa/jJwlXNufWWfoEilOAfXXOMX\nXBxf4cBASSb33ONbQq6+2s90HIt27PBz3cybBy+9BIcfHnREIoGrHa0Hcs69gy8eDt82Muz6AuD3\npdzvFSDmJhOUBPPss/Dvf8MTT0CLFkFHI7GkWTO/yOqll/pZrS+/POiIinIOrroK/u//4PHHfReV\niFRpqQaRxLRuHdx8MxxzjO+eEinu4ovh2GP9umRr1wYdTVGjR8O0aTByJFxxRdDRiMQMJTgit98O\nGzb4glINp5XSmPkZjjdv9u+XWPHUU74A+tJL46cQWiRK9Gkuye2TT3yz/tCh0L170NFILDvwQD91\nwBNPwMcfBx0NvPceXHklnHIKTJ2qle5FiqnSYpvxQottSqXk5/tizPXr/dwhaZVaVk2S2dat0LWr\nLzr+8kuoHbUyxqK++gqOOw722cfXjjUuMTWYSMKK2GKbIgnr4Yf9qJOHH1ZyI5WTlgYPPeTfN5Mn\nBxNDdjacfrovfn7nHSU3ImVQgiPJaeVKvxTDGWdA375BRyPx5OyzfYLxl7/AqlXRfewNG/woqd9+\n8xP5tWsX3ccXiSNKcCQ5DR0KO3fCI4+odkGqxsy/b/Lz4cYbo/e4v/3mk/EffoDXX/fLiYhImZTg\nSPJ55x145RX/C1yzvUp17L03DB/uV+x+//2af7yCArjkEl9v88wzcPzxNf+YInFORcaSXHJz/WiY\nevXg66/9mj0i1ZGXBwcd5Cfa++abml3U8uab/TIi99/vr4skMRUZi5Tm7rvhxx/9nCZKbmRPpKbC\nlCmwZEnNLu/x8MM+uRkyBG66qeYeRyTBKMGR5PHdd/6L6KKL4IQTgo5GEsEpp8AFF/jE+YcfIn/+\nV1/19WJ9+8KDD6peTKQKlOBIcihcTLNhQ3jggaCjkUQycaJvDRwyxL/PIuXjj2HgQDj6aHj+eahV\nK3LnFkkCSnAkOWRlwT//CffeC61aBR2NJJJ27WDMGL/Y5WuvReacixZB797QoQO88QbUrx+Z84ok\nERUZS+LbsAH239+PmPr4Y603JZGXnw+ZmX7h1j2dFfvnn6FHDz9r8ief+NmKRWQXFRmLFBoxwq8A\nrcU0pabUru3fXytXwl13Vf8827bBmWfC6tXw1ltKbkT2gD7tJbF9/jk89hhcfz0cemjQ0Ugi69ED\nrrgCJk2Cb7+t+v3z833B8n//C3//Oxx5ZORjFEkiSnAkceXnw1VXQdu2MHp00NFIMrj3XmjaFK6+\n2k/OV1mFRfBvv+2nMOjdu+ZiFEkSSnAkcT36qF91+cEHoVGjoKORZNC8uZ+KYPZsePbZyt/v7rvh\n8cfh9tvhf/6n5uITSSIqMpbElJPjC4uPOcYvSqj5QyRaCgrg2GNh8WI/Gmqvvco//tln/TIMgwb5\n63qvipRLRcaS3G68EbZv9zPN6gtDoiklxRccb9jg16sqz8yZcPnlcNJJ8OSTeq+KRJASHEk877/v\nizRHjNAoFAlG9+5www0wdSp89lnpx8ydC/36QdeufvFXLR0iElHqopLE8ttvfgHElBSYN8+vFyQS\nhC1bfDdp69Z+NF/t2rv3rVjhZyg2g08/hfT04OIUiTPqopLkkpUFnTr5GV+XLIFzzlFyI8Fq1MgX\nuH/1le+yKrRxI/Tq5Sfye/ddJTciNUQJjsS/rCwYPBiys3dve+ghv10kSOeeCz17wq23+kQmJQXa\ntPGzHb/2mm9tFJEaoQRH4seOHf7X76pVfoTKV1/BRx/5guLc3KLH5ub6GhyRIJnBqaf6rtNVq/x8\nN3l5vrvqp5+Cjk4koakGRyKroMAnF1u3+mnnC//Ku13ZY7dvr1osZlWbbE2kJnTqVLR1sVBGBixb\nFu1oROJeZWtwald0gNSQrCzfwrB8OXTsCOPGwcCB0Xls5/wvykglHuG3f/21arHUresXJmzY0P8V\nXm/duujt4vvDbw8a5BcoLK5jx8i8XiJ7Yvnyqm0XkYhQghOEwpqRwm6V7Gx/G4omOdu3R7YFpPB6\nbm7VWjZq1So9sWjWzNcVFE88KpOUFF6vHYG34IQJRV9PgAYNfNIoErSOHUtvwVECLlKjotZFZWan\nAQ8BtYAnnHP3FtvfEXgGaBo65jbn3DuhfbcDlwM7geudc++V91gx30VVVpN17dq+5aIwEcnPr/w5\nzfyXenUTjfKSlLp1Y38CsiBbxETKU/wHDfj/q1On6j0qUg2V7aKKSoJjZrWAxcApwErgC6C/c25B\n2DFTga+cc381s27AO865TqHrLwBHAu2AmcB+zrmdZT1ezCc4KSm+m6g0l11WvSSlfv3YT0JEkpUS\ncJGIibUanCOBJc65pQBmNh3oAywIO8YBjUPXmwA5oet9gOnOuTzgRzNbEjrfJ9EIvEaU1WSdkeGn\naxeRxDJwoBIakSiL1jDx9sCKsNsrQ9vCjQIGmdlK4B3guircN76MG+ebqMOpZkRERCRiopXglNZ3\nUryPpj/wtHMuHTgdeM7MUip5X8xssJnNMbM5a9as2eOAa9TAgb7/PSPDdytlZKg/XkREJIKi1UW1\nEugQdjud3V1QhS4HTgNwzn1iZvWAFpW8L865qcBU8DU4EYu8pqjJWkREpMZEqwXnC2BfM+tsZnWB\nC4E3ih2zHDgJwMy6AvWANaHjLjSzVDPrDOwLfB6luEVERCQORaUFxzmXb2ZDgPfwQ8CnOefmm9lo\nYI5z7g3gJuBxM/szvgvqUueHeM03sxfxBcn5wLXljaASERER0VINIiIiEjcqO0xci22KiIhIwlGC\nIyIiIglHCY6IiIgkHCU4IiIiknCU4IiIiEjCUYIjIiIiCUcJjoiIiCQcJTgiIiKScJTgiIiISMJR\ngiMiIiIJRwmOiIiIJBwlOCIiIpJwlOCIiIhIwlGCIyIiIglHCY6IiIgkHCU4IiIiknCU4IiI1LCs\nLOjUCVJS/GVWVtARiSQ+JTiSEPQFIrEqKwsGD4bsbHDOXw4erPeoSE1TgiNxT18gEqvWr4ebboLc\n3KLbc3P99iVL4LffgolNJNGZcy7oGCIuMzPTzZkzJ+gwJAoKCiAjA1auLLkvIwOWLYt6SJKE8vPh\n++9h7lyYN8//zZ1b+vuyNC1aQIcOkJ5e+mX79lCvXs0+B5F4YWZfOucyKzqudjSCEakq5/yv35yc\n3X8//VT0dk4OrF4NO3aUfo7sbBg1Crp1gwMOgH33hbp1o/o0JAGtXVs0iZk3D+bPh7w8v792beja\nFY4/Hrp3hwkT4JdfSp6nVSsYPx5WrPCJ0IoVPiGfPRs2bCh5fMuWFSdBqak1+tRF4ooSHIkq5/yH\nd/FkpbTb27eXvP9ee0HbttCuHey/v7987LHSvxBq14bRo/1jAtSq5ZOcbt2K/nXpol/HUtKOHfDd\nd7uTmcKE5qefdh/TurVPYq67zl927+7fl+GJRvv2vss0vJuqQQOYOBEGDiz9sbdt2530FL/88Uf4\n979h48aS92vVquwEKD1dSZAkFyU4EhHOwaZNZbe0hG8r/KUbrmlTn6y0bQvHHeevF/4VJjRt25ae\niBxwQOlfIFOnQr9+sGgRLFiw+2/+fJgxA3bu9MempMA++5RMfPbf359HEt/PP5dslVmwYHfrYN26\n/j1xyim7E5nu3X2CU5HCJGbECFi+HDp2hHHjyk5uABo29Il3ly5lH7N1q096iidAK1bADz/ArFn+\n/2RxrVtXnASppVMSgWpwpFzOwZYtFbe25OTAr7+WvH/jxiUTleLJS9u2e55IZGVV7QskLw8WLy6a\n+CxY4Lfl5/tjzPyIrMIursLEp2tXSEvbs3glGHl5vlWmeK1MeBdSu3a7E5iDD/aXXbpAnTrBxV1d\nW7aUngSFX5aVBJWWABVeb9dOSZAEp7I1OEpwktiWLZXrKtq2reR909JKb2Upvq1hw+g/rz2xY4cf\n2TJ/ftHEZ9Giol1mHTvuTngKk5+uXaFJk+Bil92c8+/d4q0y3323O4FNTYUDDyzaItO9uy/4TSaF\nSVBh609pSdDmzUXvY1Z2ElR42a5dfCaFEvuU4CRxgrNtW8XdRDk5vom7uAYNSk9Uit9u1Cj6zytI\n+fmwdGnRbq4FC/wXZvgw3/btS3Z1devma4ekZvz2m//3KF4rs27d7mM6dCjaItO9u6/Hqq1O+krZ\nvLn8VqAVK3yiFM4M2rQpPwlq27Z6SVBVW2wlsSjBifEEpzr/QXNzdycp5bW8FP+1Bb52pXjiUlry\n0qiR/2CSytm50498KV7js3Bh0ZqgNm1KT3xatgws9LjjnP8yLd4qs3jx7nqq+vXhoINKtso0axZs\n7Mlg8+byW4FWrCj5oyolxf/fKK8mqF27oolo4bxXpdXcKclJDkpwYjjBKe0/aP36cMstvsm8rOSl\ntFETqanldxEVXm/SRIlLNBUU+OS1eI3PggVFf+m2aFG0vqfwr3Xr5P73ys2Fb78t2iozb17R0XKd\nOpWsldlnHz9aTmKPc7uToLISoBUrSnaJp6T4z7LCpOe990q2FoHmvUomMZfgmNlpwENALeAJ59y9\nxfZPAk4M3WwAtHLONQ3t2wl8E9q33Dl3VnmPFesJTqdOfo6W8tSpU3F9S7t2/pdpMn8RxpvCVoji\nSc/8+UWLPZs1K1nj062b/zdPpH/vwpmni7fKfP/97uH9DRuWbJE56CDVOyWiwtGY5SVBixaVfl8z\n/8NCEl9MJThmVgtYDJwCrAS+APo75xaUcfx1wKHOuctCt7c65yo9biXWE5yUlN0f3uHM/Ad827bQ\nvHlifZFJ+QqLYktLfNav331c48ZFW3oKk58OHWL//bJ1K3zzTclWmfAu1X32KVkr07mz/z8jAr6l\nZvny0rerBSc5xNpMxkcCS5xzSwHMbDrQByg1wQH6A3dGKbao69ix9Bacjh39L1NJPma7W+dOPnn3\ndudgzZqSo7reegumTdt9XFqaH8VVvKurcAHSaCoo8JPRFW+V+eGH3cc0buyTl0GDdicyBx6YfMXr\nUnV33116Dc64ccHFJLEpWglOe2BF2O2VwFGlHWhmGUBn4MOwzfXMbA6QD9zrnHu9pgKNhnHj9B9U\nKsfMz07bqhWceGLRfWvX+mLm8FFd778Pzzyz+5j69UtPfPbeu/xalcoWwW/eXLJF5ptvdheTmvnR\nSoceCpdcsrtlJiMj9lucJDZVZ+JESU7R6qI6D+jpnLsidPsi4Ejn3HWlHDsMSA/fZ2btnHM5ZrY3\nPvE5yTn3Q7H7DQYGA3Ts2PHw7IqKXAKmYY5SUzZs2J34hCc/4Qs/pqb6yeuKFzjvsw+8+GLpCfjo\n0T4xCW+VCe8SaNq0aNdS9+7+/PE2F5KIxLZYq8HpAYxyzvUM3b4dwDl3TynHfgVc65z7uIxzPQ28\n5Zx7uazHi/UaHJEgbN5cNPEpTH7CfwsUzklS1gKm4Lu8unQpWSuTnq5WGRGpebFWg/MFsK+ZdQZW\nARcCA4ofZGZdgGbAJ2HbmgG5zrk8M2sB/B4YH5WoRRJI48Zw1FH+L9zWrX7CwsKk5777yj7HnDm+\npad+/ZqNVURkT0UlwXHO5ZvZEOA9/DDxac65+WY2GpjjnHsjdGh/YLor2qzUFfibmRUAKfganLKK\nk0WkitLSIDPT/wFMn156EXxGBhx+eHRjExGpLk30JyJFaKZYEYllle2i0uwSIlLEwIE+mSkc6ZSR\noeRGROKPlpoTkRIGDlRCIyLxTS04IiIiknCU4IiIiEjCUYIjIiIiCUcJjoiIiCQcJTgiIiKScJTg\niIiISMJRgiMiIiIJRwmOiIiIJBwlOCIiIpJwlOCIiIhIwlGCIyIiIglHCY6IiIgkHCU4IiIiknCU\n4IiIiEjCUYIjIiIiCUcJjoiIiCQcJTgiIiKScJTgiIiISMJRgiMiIiIJRwmOiIiIJBwlOCIiIpJw\nlOCIiIhIwlGCIyIiIglHCY6IiIgkHCU4IiIiknCU4IiIiEjCiVqCY2anmdkiM1tiZreVsn+SmX0d\n+ltsZhvD9l1iZt+H/i6JVswiIiISn2pH40HMrBYwBTgFWAl8YWZvOOcWFB7jnPtz2PHXAYeGru8F\n3AlkAg74MnTfDdGIXUREROJPtFpwjgSWOOeWOue2A9OBPuUc3x94IXS9J/CBc259KKn5ADitRqMV\nERGRuBatBKc9sCLs9srQthLMLAPoDHxY1fuKiIiIQPQSHCtlmyvj2AuBl51zO6tyXzMbbGZzzGzO\nmjVrqhmmiIiIJIJoJTgrgQ5ht9OBnDKOvZDd3VOVvq9zbqpzLtM5l9myZcs9DFdERETimTlXVkNK\nBB/ErDawGDgJWAV8AQxwzs0vdlwX4D2gswsFFioy/hI4LHTYf4HDnXPry3m8NUB2pJ9HDWkBrA06\niASi1zOy9HpGll7PyNLrmZwynHMVtmREZRSVcy7fzIbgk5dawDTn3HwzGw3Mcc69ETq0PzDdhWVd\nzrn1ZjYGnxQBjC4vuQndJ26acMxsjnMuM+g4EoVez8jS6xlZej0jS6+nlCcqLThSNv0HjSy9npGl\n1zOy9HpGll5PKY9mMhYREZGEowQneFODDiDB6PWMLL2ekaXXM7L0ekqZ1EUlIiIiCUctOCIiIpJw\nlOCIiIhIwlGCIwnDzEqb9VpERJKQEhxJJA2CDiBZmVmtoGMQqSz9GEoOSnBijP7jVY+Z9QbeNLMG\nZqb3dRSZWTNgn9D148ysScAhJYTSPgv03q46M2trZoeFrvc1s45Oo2uSQlRmMpbKMTMLW6LiaOAX\n4DfnXFnrdglgZj2B24CxzrlcfQlEXRfgTDNrBfwBODLgeOJesc+C3wM7gB+cc+vC90mlpAAvmNk8\noBkwMOB4JEr0RRBDwj7QbgLGAbcC48xs/0ADi2FmdhDwLvAX59y7ZtYZ+IuZ1Q84tIRX2MLgnPsU\naAwMAh5wzm0NNLAEEPZZcC1wP9AbWGBmrZTcVE7Y+3MVfr6cM4CZzrmfzayOhQQapNQoJTgxxswO\nB05xzp2E//dpACwys7rBRhZbwj6YlgGvAeebWSfgWeAX59yvwUSWHIq1MBwKPA3cCRwa6gZoGNqX\nGlyU8c3MMvGJzcn4BSUXErawpL6cy1bs/bk/MBc4GxhmZlc553aE9jcKMk6pWUpwAlbKh1Qt4Dsz\nGw6kAxeH/iP2UKtEEXUBnHNb8E3OacAPwIvOub+qm6pmhX15DMUnNjnOufuBxfgvkh6hfbfo36Jy\nSvks+AV4G7gF3/rQ0++w8TQAABEfSURBVDlXYGYDzayBWnLKVqw1/FHgO+fce/jX8R4z629mZwH3\nm1mdAEOVGqSZjANU7FdG81D/ejPgCaAT8Hvn3G9mdjVwHnC2c25TcBHHBjM7Fbga/6tsnnPu1VCL\nwWNALefcgNBxtZxzOwMMNaGFCrvvwLc4bjazlvgWhvOB4/H1OIOcc/MCDDMuFPssaIqvuUkBZgEN\nnXP7h/YNBK4ALnDO/RJQuHHBzC4A/gycGnp/dnLOLTOzA4GngK3A9c65bwINVGqMEpwYYGbXAUcB\nPwP/AtoDXYGGwHzgEmCgc+7bwIKMEWZ2GnAXviuqFdAOGO+c+97MGuF/rdUHztMv3Mgq/BI2s5RQ\nS8LVwBH4+oYzgVPxr/1h+Ba2Os65jcFFHH/M7HrgJGAb8AE+wfkM+Bu+zulY4BJ9KZdUvPjazPoC\nhwAr8J+pZ+O7+f6MTyB36gdjYlPTccDM7ELgXOBa4HTgUOAZ4ElgCbAT/2tNyY3ZXsA7+NFSU4DH\n8V+kzWFXd9Xl+Pf1/wYVZyIq9uXRLHT5v0BLYBQwDzgB36p2sHNum5KbqjGzQUBffAvNDuAc59yP\n+NawbPznwQVKbkoq1gLWKNSdvwT/+XAh8AVwJb6FsYNzbr2Sm8SnFpwoK+VXxvX4uoU2wADgrFC3\nVHPn3Lqg4oxVZnYGMB7oEWp2fgdfiP01/pfaNMCAus651cFFmpjM7H/wX8LZ+O7BR8P2nY0f/Xey\npjaoWLEv5Tr4VrDvgT/if+z0BgqAjqFERypgZjcCx+DrF8cACwpfOzPrg68XO0evZ3LQPDgBCQ3/\nXIsfBXQ3sCE0cgozuxlIM7O71M1SlHPubTMrAL40s//D/0KbAuyF/+XbFbjRObc+wDATUugLYghw\nMX7um+PNbAIwHP/lPBbfNajkpgLFkpvL8YlMYf3dbOfcaaF9g4EuZnaHRgaWz8x6Af2BnkA/4Cyg\nlZmtCl0fjh+0oeQmSSjBiRIzOwBY7JzbESqSPRnfnbIX/lfbx2Z2MP/f3r1H2z2feRx/fyIu06GC\nKknr1rh0GToNQhWRGqaClIygcQkJHaI6mlDFuIyRVqmGipZpMO24NdXphNCFutW9dUmiHWVcpohL\n0kjcchFJPvPH8z385uRyTtgnv3N2ntdaWevsvX/nd745a5/9e37f7/N9HtiamMk5IoObJSv1bkYA\ntwM9bU8DkDQOWNf2W7UOsHmtBVxte5Kkp4CngFHARkTu2Jdtv1jnALuCsm15FjBN0q7A9sBptt8s\nnxObS/o0sD9wHLEslcFN23oCU8rNzRWSpgHnA7cBk4B9bb9a5wDTipU5OCtASYwdB2wsaUtgONC9\nrAM/S9QQ+TgxkzOYSCL877rG2xXYvoPY8nmXpA3Kc4tsz1j2d6b2WEqNlVeAYyX1sT3X9mRgbWB9\n2zMyuGlbmWX4JdCjBDG3EnlMcwFsH01sLPhXYqnqENtP1jTcTqv6/tQHfdD+AHST1BfA9kTgMaCX\n7ecyuFn5ZA5OB1O0ERgDjLB9b0mU3Qc4Avil7XHluO5EDZxVbM+pbcBdTGVdfQfbi+oeT7Mp25J3\nIGoMPQhsARxLVNddCziJyGmYWtsgu4jyWXA5Ue5hcnluX2KJ9Rvlglw9flXb7634kXYdZXmvN7AA\n+E/i5vENYoZsBvHZ0N9RzTitZHIGpwOVD7RriD+0yZK6l+nTCcDVRNXX4QC2F9h+N4Ob5WP7RqBf\nBjeNV5YBRxFJ8D2J7fnvEVvxRxB5Dl/L4KZt5bPgx8SF9zMtz9u+Bfgm8CNJ+1W/J4ObZSu7zk4E\n7iRKE+wH3E0E41sBewEHZHCz8soZnA4iaSdiu/fRxFr6hkQS5jOllkgPYiZnAHCr7WtrG2xKgKTV\nbM+vPL4AuMH2I6UA5UBgR9snKFowLLS9oK7xdhWSvkgEN98C3gL+GbjJ9hWVY/YHriNmw26tZaCd\nXMuyVPn8XA24mEjIvq4U+jwHWN32N8rxa9ieV9+IU91yBqfjLCA+rB6wfQowh/hg27zsoHiDaBI5\nkbgDSak2JfH9IkmnVPIbehFF0bA9i6h107O0CXg3g5t2mwYMs/0bIr/mJ8BASce0HFBmIg8mZh9S\nKyV36UfAJaUi8XzgaaCvpJ62ZxNVtbeR1LN827s1DTd1EhngNJikLRRN8noA729HtH0cEeScAfQu\nQc4s4g4567Wk2lTyxB4iZhxPKS+NAhZIurA87k0kFa+xwgfZBUnqVv7On7M9CcDRaf1uPghyhrUc\nb/sW28/UNNxOqwTf5wGPAvOIUgUA9xJ5YPtI2oJYklqNaMHwfj+qtPLKJaoGKgmD5xJF0NYEPksU\n7ptUOeZS4s74ZNvP1zLQlApJXyCWRkbavrHs+BtCzCo+QVTS/j4wn9gOPjQr6batVZ2bXQEDk8tM\nA2VJpR8RTF5l++raBtuJSfoS0XD0bx3tWAYTO00fIHIZtyJ2U/4NccN+ku0pdY03dS5ZB6dByoXh\nTOJC8dvy3JnAjZIGtvzRlfyFHxB3IinVrTfwJNDShPAiYuvyV4AdgfG295a0ITDP2X6hXSrBzUlE\nDt504C1JVwIP2p4t6T4igMxt4Es3lQhcPkfUCzudCG56EbWX9rM9UtHodWEW+ExVOYPTAGXr9wxi\ntubmanKbpLOJZpl9nL1PUidUdku1FJy7wvaFisal5wMzbJ9V6wC7kFYzN/sBX7c9QNJoIsfmduBa\n2w+1Pj79f5JWsb2wBN73EDfkx9oeX14/B9iSKIqa+WBpMZmD0wDlrmEgcJ6ih9S8sssE2+cQ7Ri2\nqHGIKb2vdRE/25cBNxLNCSdJWtvRuPRpYENJqy6l8F+qaBXcfJ6YcTi+5Nn0JYLILYGzJe0OmSey\nJJXdUgtLkPNHoov6e0QOWIvpwEyizUVKi8klqgbxBz2Sfi9pB9uzKoW63iaXpFIn0OoifADwV8Dj\ntn+h6MB8JLCwtBM4Cjgs67G0T6uZmzOJ7urvAtsQeTbTJd1RHmel8iVYwvvzY5Ketv2YpD2AuyUt\nIG4ajyJ2p2WAk5YoA5wGcvRIOgF4tBLkDCVq4EyveXgpVS/CI4lCfQ8BQyVda/tnkhYCZxHv2cHZ\nJqB9Wi7MkvYiOqoPd+kfJekPwGhJfYDdiaAxW4osQeX9eQJR7f2nwD2SDi9J8P2J5PeZwG62/1TX\nWFPnlwFOg1WCnHsl/Zj4Iz3adgY4qTaSNiKuH1Ml7Qzsans3SacRwczu5Rp9jaTXgaecXZfbpOiD\nNtPRRHcnYgflekQ7i38sh40HZhO7pobn7snFtZq52ZrYybcn0bn+BeBiSX9divptTvTy+5/6Rpy6\ngkwy7iBlmvpXRHJxTken2khaD7iEqCNyHdHYcW1gJ6LlwkBi1qE/cJHt/6hnpF2PpIHAyUQO057E\nFuZNgauA22yfXTm2Wy6nLK5VcDMEuIsolNqfSNLeQ9Io4EJgn6z0nNork4w7iO2bgR4Z3KQ6lYvH\n68BYYFvgIKCb7ZeATYA7Hf3PniSaaebFox0qibATify684naVnPKst4wYlZsTMv3ZHCzuJJE3BLc\nDAKOIRoOvw6sQ3QIB3iRaKb5VC0DTV1SLlF1IGfjzFSzyi6dTYBPEfVt1ij1WO4F7pO0GbALUeYg\nl1LboXJR3ha4mcixu0HSbrZn2v6TpOOB70ha3/Zf6hxvZ1QKIG4p6ffEjM0JRN2lVyStQgQ1u0sa\nD2wGHGT7hfpGnLqanMFJqcmVZZRvAQcC/0R0Xj7S9iPEMtVkYH/bz9Y3yq6hul2+bAU/w/ZY20cA\n9xFJ2y2/877AwRncLK4URh1LFDpcF1iFKFMwRNLnbC8kfp+XEa0thmZwk5ZX5uCk1GQkbWB7WuXx\ncODvbB9WHg8ALgeuBMY6eqKl5VRmGX4OfNelHYuknwF9iNosh+YutMWVGkBXErvJfld5fjCRoL0N\nMM72EzUNMTWJnMFJqYmU+jWvShojqWUXz+PA25K+CLHTjyhzvxHRIym1QdL2ir5dSDpG0mXApURz\nx40rhT2PBL4KDMjgZqn6EIF1Nbj5HpEI35dIhh9Z3sspfWiZg5NSc5lNLJNMAw4sW5fvIm5mdpHU\nr7y2AfC17C3VtjLjNZpIHAZ4mMi56U1UJ/434DZJnyBaMYzNhOLFVXZL9QberDw/gOgtNQi4BngV\nmALkezN9JBngpNREbL9Ukja3I7osHwL8PXFnPJ3YSfUaMMr2i7UNtIsouSJnAKfZfqIEMbNs31Re\nnwvsAYwC9gbuzuBmySoJ7xOAUyVtZ/tx4A5iN998SVcRLS5udunnl9KHlUtUKTWJSgLst4mlp08A\nLxP1RG4HtiYuHmdk+YK2KZro/hr4vu3bS4G5/wI2rxz2CPCO7ddtX2v7lTrG2sU8THQE/6qkHW2/\nV4KbIcQuv8cyuEmNkDM4KTWJ0ipAgIgdKWOImZwTbU8oOQ2v5bJU+9ieWXZDnSvpeeAHwATbv60c\n9iawlaQewFs5e9M227MljQOOBi6QNIkoPjkYOCAraKdGyV1UKTUhSVsR22zH2j637vF0ZWWZ6tfA\n6ba/V4rTLSy5I3OAR23PrneUXU9p7rodsBcx03iP7WfqHVVqJhngpNSkJA0jCvxdkEUnP5rSRHMs\n8AXbb0g6CjgeOCRnHFLqnDIHJ6Xm9RCwfd2DaAa2fwOMJCo/jyCWV4ZlcJNS55UzOCk1MUkfy9mb\nxskmuil1HRngpJTScsigMaWuIQOclFJKKTWdzMFJKaWUUtPJACellFJKTScDnJRSSik1nQxwUkop\npdR0MsBJKa30JHWXZEmbNuh8UyX1b8S5UkofTgY4KaV2kfRO5d8iSXMrjw+re3zLQ9L9pRpxSqlJ\nZbPNlFK72F6z5WtJfwaOsX3H0o6X1N32ghUxtpRSai1ncFJKDSFptKTxkq6X9DZwuKRukk6X9Jyk\nGZJ+LmmdyvfsIulhSW9Imiyp3zLOP1XSyZL+WGaNfiJpA0m3SXpL0u2lq/cyzy3pfGBn4PJynosr\nP+bLkp6VNEvSJZVzdZN0lqQXJE2X9FNJH6+8flR5bYakUxvzG00pfRQZ4KSUGmkQcB2wNjAeGAXs\nC/QDPg3MBi4BkLQRcBNwNrAucCrwK0nrLeP8/wDsAXwWOBC4BTgF+CSwOvD1ts5t+9tEn67jbK9p\n+5uV8+9D9O/qQwRoe5bnjwEOB/oDvYF1gB+Wn7UtcClwKPApoBewYft/ZSmljpABTkqpke63PdH2\nIttzgWOB022/bHse8C/AwZK6AUOBm2zfVo6/FZgC7L2M8//Q9nTbU4H7gYdsTynnnkAEJnzIcwOc\nZ/tN238G7gE+X54/DLjQ9v/afhs4HTi0/D8OAibYfsD2u+U1tfP3lVLqIJmDk1JqpJdaPd4YmChp\nUeU5EzMumwBDJA2qvLYqcOsyzj+t8vXcJTxuyRP6MOcGeK3y9ZzK+XoBL1ReewFYDVi/vPb+/9v2\nO5JmtvFzUkodLAOclFIjtW5uNxU41PbvWh8o6SXg322P6IBxtHXu5W3C9woRNLXYGJgP/AV4Fdis\n5QVJaxLLYimlGuUSVUqpI10OfFfSxgCSPinpK+W1q4FBkvaStIqkNSR9SVKvBvzcts49DfjMcpzv\nemCUpE0lrQV8B7je9iLgBmB/STtLWh0YzfIHUCmlBssAJ6XUkcYQy0J3lp1VDwJ9AUqeyyDgTGIm\n5EXgJBrwudSOc19MLGG9IWlMO045jkiavg94HngbOLH8rCfK178AXiaWuV5b8mlSSiuK7LzRSCml\nlFJzyRmclFJKKTWdDHBSSiml1HQywEkppZRS08kAJ6WUUkpNJwOclFJKKTWdDHBSSiml1HQywEkp\npZRS08kAJ6WUUkpN5/8Ava6gLs7HRWcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20c140e1748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_score_tree = train_score\n",
    "cv_score_tree = cv_score\n",
    "\n",
    "ymin = np.min(cv_score)-0.05\n",
    "ymax = np.max(train_score)+0.05\n",
    "\n",
    "x_ticks = clf_tree.keys()\n",
    "\n",
    "plt.figure(figsize=(8,5))\n",
    "plt.plot(range(len(x_ticks)), train_score_tree, 'ro-', label = 'training')\n",
    "plt.plot(range(len(x_ticks)),cv_score_tree, 'bo-', label = 'Cross-validation')\n",
    "\n",
    "plt.xticks(range(len(x_ticks)),x_ticks,rotation = 45, fontsize = 10)\n",
    "plt.xlabel(\"Tree method\", fontsize = 12)\n",
    "plt.ylabel(\"Score\", fontsize = 12)\n",
    "plt.xlim(-0.5, 5.5)\n",
    "plt.ylim(ymin, ymax)\n",
    "\n",
    "plt.legend(loc = 'best', fontsize = 12)\n",
    "plt.title(\"Different tree methods\")\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}