贝叶斯 - 文本分析与关键词提取.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 如何安装jieba：https://pypi.python.org/pypi/jieba/ ###"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import jieba #需要先安装jieba库"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.chdir('/Users/a1/Desktop/算法实战/贝叶斯_新闻分类/贝叶斯-新闻分类/data')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1）导入数据&处理包"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>category</th>\n",
       "      <th>theme</th>\n",
       "      <th>URL</th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>汽车</td>\n",
       "      <td>新辉腾　４．２　Ｖ８　４座加长Ｉｎｄｉｖｉｄｕａｌ版２０１１款　最新报价</td>\n",
       "      <td>http://auto.data.people.com.cn/model_15782/</td>\n",
       "      <td>经销商　电话　试驾／订车Ｕ憬杭州滨江区江陵路１７８０号４００８－１１２２３３转５８６４＃保常...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>汽车</td>\n",
       "      <td>９１８　Ｓｐｙｄｅｒ概念车</td>\n",
       "      <td>http://auto.data.people.com.cn/prdview_165423....</td>\n",
       "      <td>呼叫热线　４００８－１００－３００　服务邮箱　ｋｆ＠ｐｅｏｐｌｅｄａｉｌｙ．ｃｏｍ．ｃｎ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>汽车</td>\n",
       "      <td>日内瓦亮相　ＭＩＮＩ性能版／概念车－１．６Ｔ引擎</td>\n",
       "      <td>http://auto.data.people.com.cn/news/story_5249...</td>\n",
       "      <td>ＭＩＮＩ品牌在二月曾经公布了最新的ＭＩＮＩ新概念车Ｃｌｕｂｖａｎ效果图，不过现在在日内瓦车展...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>汽车</td>\n",
       "      <td>清仓大甩卖一汽夏利Ｎ５威志Ｖ２低至３．３９万</td>\n",
       "      <td>http://auto.data.people.com.cn/news/story_6144...</td>\n",
       "      <td>清仓大甩卖！一汽夏利Ｎ５、威志Ｖ２低至３．３９万＝日，启新中国一汽强势推出一汽夏利Ｎ５、威志...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>汽车</td>\n",
       "      <td>大众敞篷家族新成员　高尔夫敞篷版实拍</td>\n",
       "      <td>http://auto.data.people.com.cn/news/story_5686...</td>\n",
       "      <td>在今年３月的日内瓦车展上，我们见到了高尔夫家族的新成员，高尔夫敞篷版，这款全新敞篷车受到了众...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  category                                 theme  \\\n",
       "0       汽车  新辉腾　４．２　Ｖ８　４座加长Ｉｎｄｉｖｉｄｕａｌ版２０１１款　最新报价   \n",
       "1       汽车                         ９１８　Ｓｐｙｄｅｒ概念车   \n",
       "2       汽车              日内瓦亮相　ＭＩＮＩ性能版／概念车－１．６Ｔ引擎   \n",
       "3       汽车                清仓大甩卖一汽夏利Ｎ５威志Ｖ２低至３．３９万   \n",
       "4       汽车                    大众敞篷家族新成员　高尔夫敞篷版实拍   \n",
       "\n",
       "                                                 URL  \\\n",
       "0        http://auto.data.people.com.cn/model_15782/   \n",
       "1  http://auto.data.people.com.cn/prdview_165423....   \n",
       "2  http://auto.data.people.com.cn/news/story_5249...   \n",
       "3  http://auto.data.people.com.cn/news/story_6144...   \n",
       "4  http://auto.data.people.com.cn/news/story_5686...   \n",
       "\n",
       "                                             content  \n",
       "0  经销商　电话　试驾／订车Ｕ憬杭州滨江区江陵路１７８０号４００８－１１２２３３转５８６４＃保常...  \n",
       "1       呼叫热线　４００８－１００－３００　服务邮箱　ｋｆ＠ｐｅｏｐｌｅｄａｉｌｙ．ｃｏｍ．ｃｎ  \n",
       "2  ＭＩＮＩ品牌在二月曾经公布了最新的ＭＩＮＩ新概念车Ｃｌｕｂｖａｎ效果图，不过现在在日内瓦车展...  \n",
       "3  清仓大甩卖！一汽夏利Ｎ５、威志Ｖ２低至３．３９万＝日，启新中国一汽强势推出一汽夏利Ｎ５、威志...  \n",
       "4  在今年３月的日内瓦车展上，我们见到了高尔夫家族的新成员，高尔夫敞篷版，这款全新敞篷车受到了众...  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#这里使用的是pd.read_table: http://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_table.html\n",
    "df_news = pd.read_table('val.txt', names=['category', 'theme', 'URL', 'content'], encoding = 'utf-8')\n",
    "df_news = df_news.dropna()\n",
    "df_news.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df_newss = pd.read_table('val.txt')\n",
    "#df_newss "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5000, 4)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_news.shape #看一下表格大概结构"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2）使用jieba对每一条content内容分词操作，存入list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "阿里巴巴集团昨日宣布，将在集团管理层面设立首席数据官岗位（Ｃｈｉｅｆ　Ｄａｔａ　Ｏｆｆｉｃｅｒ），阿里巴巴Ｂ２Ｂ公司ＣＥＯ陆兆禧将会出任上述职务，向集团ＣＥＯ马云直接汇报。＞菹ぃ和６月初的首席风险官职务任命相同，首席数据官亦为阿里巴巴集团在完成与雅虎股权谈判，推进“ｏｎｅ　ｃｏｍｐａｎｙ”目标后，在集团决策层面新增的管理岗位。０⒗锛团昨日表示，“变成一家真正意义上的数据公司”已是战略共识。记者刘夏\n"
     ]
    }
   ],
   "source": [
    "content = df_news.content.values.tolist() #在pandas中拿某一列的值，并把它转化为list格式\n",
    "print(content[1000])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "#【进行分词操作】\n",
    "content_S = []\n",
    "for line in content: #对每一个content中内容循环，分词操作\n",
    "    current_segment = jieba.lcut(line) #使用jieba.lcut()命令进行分词\n",
    "    if len(current_segment) > 1 and current_segment != '\\r\\r': #挑出content中长度大于1，且不是分隔符的字符\n",
    "        content_S.append(current_segment) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['阿里巴巴',\n",
       " '集团',\n",
       " '昨日',\n",
       " '宣布',\n",
       " '，',\n",
       " '将',\n",
       " '在',\n",
       " '集团',\n",
       " '管理',\n",
       " '层面',\n",
       " '设立',\n",
       " '首席',\n",
       " '数据',\n",
       " '官',\n",
       " '岗位',\n",
       " '（',\n",
       " 'Ｃ',\n",
       " 'ｈ',\n",
       " 'ｉ',\n",
       " 'ｅ',\n",
       " 'ｆ',\n",
       " '\\u3000',\n",
       " 'Ｄ',\n",
       " 'ａ',\n",
       " 'ｔ',\n",
       " 'ａ',\n",
       " '\\u3000',\n",
       " 'Ｏ',\n",
       " 'ｆ',\n",
       " 'ｆ',\n",
       " 'ｉ',\n",
       " 'ｃ',\n",
       " 'ｅ',\n",
       " 'ｒ',\n",
       " '）',\n",
       " '，',\n",
       " '阿里巴巴',\n",
       " 'Ｂ',\n",
       " '２',\n",
       " 'Ｂ',\n",
       " '公司',\n",
       " 'Ｃ',\n",
       " 'Ｅ',\n",
       " 'Ｏ',\n",
       " '陆兆禧',\n",
       " '将',\n",
       " '会',\n",
       " '出任',\n",
       " '上述',\n",
       " '职务',\n",
       " '，',\n",
       " '向',\n",
       " '集团',\n",
       " 'Ｃ',\n",
       " 'Ｅ',\n",
       " 'Ｏ',\n",
       " '马云',\n",
       " '直接',\n",
       " '汇报',\n",
       " '。',\n",
       " '＞',\n",
       " '菹',\n",
       " 'ぃ',\n",
       " '和',\n",
       " '６',\n",
       " '月初',\n",
       " '的',\n",
       " '首席',\n",
       " '风险',\n",
       " '官',\n",
       " '职务',\n",
       " '任命',\n",
       " '相同',\n",
       " '，',\n",
       " '首席',\n",
       " '数据',\n",
       " '官亦为',\n",
       " '阿里巴巴',\n",
       " '集团',\n",
       " '在',\n",
       " '完成',\n",
       " '与',\n",
       " '雅虎',\n",
       " '股权',\n",
       " '谈判',\n",
       " '，',\n",
       " '推进',\n",
       " '“',\n",
       " 'ｏ',\n",
       " 'ｎ',\n",
       " 'ｅ',\n",
       " '\\u3000',\n",
       " 'ｃ',\n",
       " 'ｏ',\n",
       " 'ｍ',\n",
       " 'ｐ',\n",
       " 'ａ',\n",
       " 'ｎ',\n",
       " 'ｙ',\n",
       " '”',\n",
       " '目标',\n",
       " '后',\n",
       " '，',\n",
       " '在',\n",
       " '集团',\n",
       " '决策',\n",
       " '层面',\n",
       " '新增',\n",
       " '的',\n",
       " '管理',\n",
       " '岗位',\n",
       " '。',\n",
       " '０',\n",
       " '⒗',\n",
       " '锛',\n",
       " '团',\n",
       " '昨日',\n",
       " '表示',\n",
       " '，',\n",
       " '“',\n",
       " '变成',\n",
       " '一家',\n",
       " '真正',\n",
       " '意义',\n",
       " '上',\n",
       " '的',\n",
       " '数据',\n",
       " '公司',\n",
       " '”',\n",
       " '已',\n",
       " '是',\n",
       " '战略',\n",
       " '共识',\n",
       " '。',\n",
       " '记者',\n",
       " '刘夏']"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "content_S[1000] #查看分词结果"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3）把分词后content的list，创建对应的dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content_S</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[经销商, 　, 电话, 　, 试驾, ／, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[呼叫, 热线, 　, ４, ０, ０, ８, －, １, ０, ０, －, ３, ０, ０...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 在, 二月, 曾经, 公布, 了, 最新, 的, Ｍ, Ｉ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[清仓, 大, 甩卖, ！, 一汽, 夏利, Ｎ, ５, 、, 威志, Ｖ, ２, 低至, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[在, 今年, ３, 月, 的, 日内瓦, 车展, 上, ，, 我们, 见到, 了, 高尔夫...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           content_S\n",
       "0  [经销商, 　, 电话, 　, 试驾, ／, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, ...\n",
       "1  [呼叫, 热线, 　, ４, ０, ０, ８, －, １, ０, ０, －, ３, ０, ０...\n",
       "2  [Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 在, 二月, 曾经, 公布, 了, 最新, 的, Ｍ, Ｉ...\n",
       "3  [清仓, 大, 甩卖, ！, 一汽, 夏利, Ｎ, ５, 、, 威志, Ｖ, ２, 低至, ...\n",
       "4  [在, 今年, ３, 月, 的, 日内瓦, 车展, 上, ，, 我们, 见到, 了, 高尔夫..."
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_content = pd.DataFrame({'content_S':content_S})\n",
    "df_content.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4) 开始清洗：先去除停用词"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "#os.chdir('/Users/a1/Desktop/算法实战/贝叶斯_新闻分类/贝叶斯-新闻分类/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>stopword</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>!</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>\"</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>#</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>$</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>&amp;</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>'</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>*</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>,</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>--</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>..</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>......</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>...................</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>./</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>.一</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>记者</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>数</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>年</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>月</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>日</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>时</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>分</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>秒</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>/</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2583</th>\n",
       "      <td>80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2584</th>\n",
       "      <td>81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2585</th>\n",
       "      <td>82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2586</th>\n",
       "      <td>83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2587</th>\n",
       "      <td>84</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2588</th>\n",
       "      <td>85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2589</th>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2590</th>\n",
       "      <td>87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2591</th>\n",
       "      <td>88</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2592</th>\n",
       "      <td>89</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2593</th>\n",
       "      <td>90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2594</th>\n",
       "      <td>91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2595</th>\n",
       "      <td>92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2596</th>\n",
       "      <td>93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2597</th>\n",
       "      <td>94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2598</th>\n",
       "      <td>95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2599</th>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2600</th>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2601</th>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2602</th>\n",
       "      <td>99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2603</th>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2604</th>\n",
       "      <td>01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2605</th>\n",
       "      <td>02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2606</th>\n",
       "      <td>03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2607</th>\n",
       "      <td>04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2608</th>\n",
       "      <td>05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2609</th>\n",
       "      <td>06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2610</th>\n",
       "      <td>07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2611</th>\n",
       "      <td>08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2612</th>\n",
       "      <td>09</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2613 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 stopword\n",
       "0                       !\n",
       "1                       \"\n",
       "2                       #\n",
       "3                       $\n",
       "4                       %\n",
       "5                       &\n",
       "6                       '\n",
       "7                       (\n",
       "8                       )\n",
       "9                       *\n",
       "10                      +\n",
       "11                      ,\n",
       "12                      -\n",
       "13                     --\n",
       "14                      .\n",
       "15                     ..\n",
       "16                    ...\n",
       "17                 ......\n",
       "18    ...................\n",
       "19                     ./\n",
       "20                     .一\n",
       "21                     记者\n",
       "22                      数\n",
       "23                      年\n",
       "24                      月\n",
       "25                      日\n",
       "26                      时\n",
       "27                      分\n",
       "28                      秒\n",
       "29                      /\n",
       "...                   ...\n",
       "2583                   80\n",
       "2584                   81\n",
       "2585                   82\n",
       "2586                   83\n",
       "2587                   84\n",
       "2588                   85\n",
       "2589                   86\n",
       "2590                   87\n",
       "2591                   88\n",
       "2592                   89\n",
       "2593                   90\n",
       "2594                   91\n",
       "2595                   92\n",
       "2596                   93\n",
       "2597                   94\n",
       "2598                   95\n",
       "2599                   96\n",
       "2600                   97\n",
       "2601                   98\n",
       "2602                   99\n",
       "2603                  100\n",
       "2604                   01\n",
       "2605                   02\n",
       "2606                   03\n",
       "2607                   04\n",
       "2608                   05\n",
       "2609                   06\n",
       "2610                   07\n",
       "2611                   08\n",
       "2612                   09\n",
       "\n",
       "[2613 rows x 1 columns]"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 取出停用词表\n",
    "stopwords=pd.read_csv(\"stopwords.txt\",index_col=False,sep=\"\\t\",quoting=3,names=['stopword'], encoding='utf-8')\n",
    "stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "pandas.core.frame.DataFrame"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(stopwords)\n",
    "type(df_content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "#定义一个函数，输入：list格式的文档与停用词表， 输出：去除停用词的文档，与去除的文档词\n",
    "def drop_stopwords(contents,stopwords):\n",
    "    contents_clean = []\n",
    "    all_words = []\n",
    "    for line in contents:\n",
    "        line_clean = []\n",
    "        for word in line:\n",
    "            if word in stopwords:\n",
    "                continue\n",
    "            line_clean.append(word)\n",
    "            all_words.append(str(word))\n",
    "        contents_clean.append(line_clean)\n",
    "    return contents_clean,all_words\n",
    "    #print (contents_clean)\n",
    "        \n",
    "\n",
    "contents = df_content.content_S.values.tolist()    \n",
    "stopwords = stopwords.stopword.values.tolist()\n",
    "contents_clean,all_words = drop_stopwords(contents,stopwords)\n",
    "\n",
    "#df_content.content_S.isin(stopwords.stopword)\n",
    "#df_content=df_content[~df_content.content_S.isin(stopwords.stopword)]\n",
    "#df_content.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5) 清洗完成第一步，去除了contents中在停词表出现过的词汇"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>contents_clean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[经销商, 电话, 试驾, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, 路, 号, 转, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 二月, 公布, 最新, Ｍ, Ｉ, Ｎ, Ｉ, 新, 概念...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[清仓, 甩卖, 一汽, 夏利, Ｎ, 威志, Ｖ, 低至, 万, 启新, 中国, 一汽, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[日内瓦, 车展, 见到, 高尔夫, 家族, 新, 成员, 高尔夫, 敞篷版, 款, 全新,...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                      contents_clean\n",
       "0  [经销商, 电话, 试驾, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, 路, 号, 转, ...\n",
       "1  [呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...\n",
       "2  [Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 二月, 公布, 最新, Ｍ, Ｉ, Ｎ, Ｉ, 新, 概念...\n",
       "3  [清仓, 甩卖, 一汽, 夏利, Ｎ, 威志, Ｖ, 低至, 万, 启新, 中国, 一汽, ...\n",
       "4  [日内瓦, 车展, 见到, 高尔夫, 家族, 新, 成员, 高尔夫, 敞篷版, 款, 全新,..."
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_content = pd.DataFrame({'contents_clean' : contents_clean}) #把处理完成的list，改成字典格式，传入DataFrame格式\n",
    "df_content.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>all_words</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>经销商</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>电话</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>试驾</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>订车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ｕ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>憬</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>杭州</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>滨江区</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>江陵</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>路</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>号</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>转</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>保常</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>叮</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>万</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>阒</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>菔</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>邪</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>自魄</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>白云</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>大道北</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>号</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>广州市</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>天河区</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>黄埔</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>大道</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>西</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>号</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>富力</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>盈泰</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993898</th>\n",
       "      <td>脸型</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993899</th>\n",
       "      <td>采用</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993900</th>\n",
       "      <td>演艺人</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993901</th>\n",
       "      <td>脸型</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993902</th>\n",
       "      <td>黑人</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993903</th>\n",
       "      <td>名</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993904</th>\n",
       "      <td>白人</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993905</th>\n",
       "      <td>名</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993906</th>\n",
       "      <td>中国</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993907</th>\n",
       "      <td>名</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993908</th>\n",
       "      <td>日本</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993909</th>\n",
       "      <td>名</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993910</th>\n",
       "      <td>名</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993911</th>\n",
       "      <td>脸型</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993912</th>\n",
       "      <td>Ｍ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993913</th>\n",
       "      <td>贾凶</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993914</th>\n",
       "      <td>宰笙</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993915</th>\n",
       "      <td>蛴</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993916</th>\n",
       "      <td>乙来</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993917</th>\n",
       "      <td>挝</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993918</th>\n",
       "      <td>黑人</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993919</th>\n",
       "      <td>美女</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993920</th>\n",
       "      <td>白人</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993921</th>\n",
       "      <td>美女</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993922</th>\n",
       "      <td>中国</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993923</th>\n",
       "      <td>美女</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993924</th>\n",
       "      <td>日本</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993925</th>\n",
       "      <td>美女</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993926</th>\n",
       "      <td>韩国</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>993927</th>\n",
       "      <td>美女</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>993928 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       all_words\n",
       "0            经销商\n",
       "1             电话\n",
       "2             试驾\n",
       "3             订车\n",
       "4              Ｕ\n",
       "5              憬\n",
       "6             杭州\n",
       "7            滨江区\n",
       "8             江陵\n",
       "9              路\n",
       "10             号\n",
       "11             转\n",
       "12            保常\n",
       "13             叮\n",
       "14             万\n",
       "15             阒\n",
       "16             菔\n",
       "17             邪\n",
       "18            自魄\n",
       "19            白云\n",
       "20           大道北\n",
       "21             号\n",
       "22           广州市\n",
       "23           天河区\n",
       "24            黄埔\n",
       "25            大道\n",
       "26             西\n",
       "27             号\n",
       "28            富力\n",
       "29            盈泰\n",
       "...          ...\n",
       "993898        脸型\n",
       "993899        采用\n",
       "993900       演艺人\n",
       "993901        脸型\n",
       "993902        黑人\n",
       "993903         名\n",
       "993904        白人\n",
       "993905         名\n",
       "993906        中国\n",
       "993907         名\n",
       "993908        日本\n",
       "993909         名\n",
       "993910         名\n",
       "993911        脸型\n",
       "993912         Ｍ\n",
       "993913        贾凶\n",
       "993914        宰笙\n",
       "993915         蛴\n",
       "993916        乙来\n",
       "993917         挝\n",
       "993918        黑人\n",
       "993919        美女\n",
       "993920        白人\n",
       "993921        美女\n",
       "993922        中国\n",
       "993923        美女\n",
       "993924        日本\n",
       "993925        美女\n",
       "993926        韩国\n",
       "993927        美女\n",
       "\n",
       "[993928 rows x 1 columns]"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_all_words = pd.DataFrame({'all_words':all_words}) #把处理完成的所有的all_words打印出来\n",
    "df_all_words"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6）计算词频，挑出出现次数最多的那几个词"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/a1/anaconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: FutureWarning: using a dict on a Series for aggregation\n",
      "is deprecated and will be removed in a future version\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>all_words</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4077</th>\n",
       "      <td>中</td>\n",
       "      <td>5199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4209</th>\n",
       "      <td>中国</td>\n",
       "      <td>3115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88255</th>\n",
       "      <td>说</td>\n",
       "      <td>3055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104747</th>\n",
       "      <td>Ｓ</td>\n",
       "      <td>2646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1373</th>\n",
       "      <td>万</td>\n",
       "      <td>2390</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       all_words  count\n",
       "4077           中   5199\n",
       "4209          中国   3115\n",
       "88255          说   3055\n",
       "104747         Ｓ   2646\n",
       "1373           万   2390"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#计算词频groupby: http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.groupby.html\n",
    "import numpy\n",
    "words_count = df_all_words.groupby(by=['all_words'])['all_words'].agg({\"count\":numpy.size})\n",
    "words_count = words_count.reset_index().sort_values(by=[\"count\"], ascending = False)\n",
    "words_count.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7) 使用词云wordcloud展示"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "#使用词云\n",
    "from wordcloud import WordCloud #导入安装的词云库\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import matplotlib\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "#https://matplotlib.org/users/customizing.html\n",
    "matplotlib.rcParams['figure.figsize'] = [10.0, 8.0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1554aa668>"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1353b2748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "wordcloud = WordCloud(font_path='simhei.ttf', background_color = \"white\", max_font_size = 80)\n",
    "word_frequence = {x[0]:x[1] for x in words_count.head(100).values}  #前100个词，画出来\n",
    "wordcloud = wordcloud.fit_words(word_frequence)\n",
    "plt.imshow(wordcloud)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 8) 使用TF-IDF 提取关键字"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "import jieba.analyse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "法国ＶＳ西班牙、里贝里ＶＳ哈维，北京时间６月２４日凌晨一场的大战举世瞩目，而这场胜利不仅仅关乎两支顶级强队的命运，同时也是他们背后的球衣赞助商耐克和阿迪达斯之间的一次角逐。Ｔ谌胙”窘炫分薇的１６支球队之中，阿迪达斯和耐克的势力范围也是几乎旗鼓相当：其中有５家球衣由耐克提供，而阿迪达斯则赞助了６家，此外茵宝有３家，而剩下的两家则由彪马赞助。而当比赛进行到现在，率先挺进四强的两支球队分别被耐克支持的葡萄牙和阿迪达斯支持的德国占据，而由于最后一场１／４决赛是茵宝（英格兰）和彪马（意大利）的对决，这也意味着明天凌晨西班牙同法国这场阿迪达斯和耐克在１／４决赛的唯一一次直接交手将直接决定两家体育巨头在此次欧洲杯上的胜负。８据评估，在２０１２年足球商品的销售额能总共超过４０亿欧元，而单单是不足一个月的欧洲杯就有高达５亿的销售额，也就是说在欧洲杯期间将有７００万件球衣被抢购一空。根据市场评估，两大巨头阿迪达斯和耐克的市场占有率也是并驾齐驱，其中前者占据３８％，而后者占据３６％。体育权利顾问奥利弗－米歇尔在接受《队报》采访时说：“欧洲杯是耐克通过法国翻身的一个绝佳机会！”Ｃ仔尔接着谈到两大赞助商的经营策略：“竞技体育的成功会燃起球衣购买的热情，不过即便是水平相当，不同国家之间的欧洲杯效应却存在不同。在德国就很出色，大约１／４的德国人通过电视观看了比赛，而在西班牙效果则差很多，由于民族主义高涨的加泰罗尼亚地区只关注巴萨和巴萨的球衣，他们对西班牙国家队根本没什么兴趣。”因此尽管西班牙接连拿下欧洲杯和世界杯，但是阿迪达斯只为西班牙足协支付每年２６００万的赞助费＃相比之下尽管最近两届大赛表现糟糕法国足协将从耐克手中每年可以得到４０００万欧元。米歇尔解释道：“法国创纪录的４０００万欧元赞助费得益于阿迪达斯和耐克竞逐未来１５年欧洲市场的竞争。耐克需要笼络一个大国来打赢这场欧洲大陆的战争，而尽管德国拿到的赞助费并不太高，但是他们却显然牢牢掌握在民族品牌阿迪达斯手中。从长期投资来看，耐克给法国的赞助并不算过高。”\n"
     ]
    }
   ],
   "source": [
    "#打印未清洗数据\n",
    "index = 2400\n",
    "print(df_news['content'][index])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['法国', 'Ｖ', 'Ｓ', '西班牙', '里', '贝里', 'Ｖ', 'Ｓ', '哈维', '北京', '时间', '凌晨', '一场', '大战', '举世瞩目', '这场', '胜利', '关乎', '两支', '顶级', '强队', '命运', '背后', '球衣', '赞助商', '耐克', '阿迪达斯', '之间', '角逐', 'Ｔ', '谌', '胙', '窘炫分', '薇', '支', '球队', '之中', '阿迪达斯', '耐克', '势力范围', '旗鼓相当', '家', '球衣', '耐克', '提供', '阿迪达斯', '赞助', '家', '茵宝', '家', '剩下', '两家', '彪马', '赞助', '比赛', '率先', '挺进', '四强', '两支', '球队', '耐克', '支持', '葡萄牙', '阿迪达斯', '支持', '德国', '占据', '一场', '决赛', '是茵宝', '英格兰', '彪马', '意大利', '对决', '意味着', '明天', '凌晨', '西班牙', '法国', '这场', '阿迪达斯', '耐克', '决赛', '唯一', '交手', '两家', '体育', '巨头', '欧洲杯', '胜负', '评估', '足球', '商品', '销售额', '总共', '超过', '亿', '欧元', '单单是', '一个月', '欧洲杯', '高达', '亿', '销售额', '欧洲杯', '期间', '万件', '球衣', '抢购一空', '市场', '评估', '两大', '巨头', '阿迪达斯', '耐克', '市场占有率', '并驾齐驱', '占据', '占据', '体育', '权利', '顾问', '奥利弗', '米歇尔', '接受', '队报', '采访', '时说', '欧洲杯', '耐克', '法国', '翻身', '绝佳', '机会', 'Ｃ', '仔尔', '谈到', '两大', '赞助商', '经营策略', '竞技', '体育', '成功', '燃起', '球衣', '购买', '热情', '水平', '国家', '之间', '欧洲杯', '效应', '德国', '出色', '德国人', '电视', '观看', '比赛', '西班牙', '效果', '差', '民族主义', '高涨', '加泰罗尼亚地区', '关注', '巴萨', '巴萨', '球衣', '西班牙', '国家队', '没什么', '兴趣', '西班牙', '接连', '拿下', '欧洲杯', '世界杯', '阿迪达斯', '西班牙', '足协', '支付', '万', '赞助费', '相比之下', '两届', '大赛', '表现', '糟糕', '法国', '足协', '耐克', '手中', '万', '欧元', '米歇尔', '解释', '道', '法国', '创纪录', '万', '欧元', '赞助费', '得益于', '阿迪达斯', '耐克', '竞逐', '未来', '欧洲', '市场', '竞争', '耐克', '笼络', '大国', '打赢', '这场', '欧洲', '大陆', '战争', '德国', '拿到', '赞助费', '不太高', '民族', '品牌', '阿迪达斯', '手中', '长期投资', '耐克', '法国', '赞助', '不算', '过高']\n",
      "--------------------------------------------------\n",
      "耐克 阿迪达斯 欧洲杯 球衣 西班牙\n"
     ]
    }
   ],
   "source": [
    "#打印清洗过后的数据\n",
    "#使用jieba，连接起来，提取关键词\n",
    "print(df_content['contents_clean'][index])\n",
    "print(\"--------------------------------------------------\")\n",
    "content_S_str = \"\".join(contents_clean[index])\n",
    "print(\" \".join(jieba.analyse.extract_tags(content_S_str, topK=5, withWeight=False)))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 9) LDA 主题模型\n",
    "    格式要求：list of list 形式，分词好的整个语料\n",
    "    找出文章中主题"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "#类似于无监督模型\n",
    "#导入gensim库，自然语言处理，参考官网：https://radimrehurek.com/gensim/\n",
    "from gensim import corpora, models, similarities\n",
    "import gensim\n",
    "#http://radimrehurek.com/gensim/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "#首先做映射表（字典），相当于词袋，\n",
    "dictionary = corpora.Dictionary(contents_clean)\n",
    "corpus = [dictionary.doc2bow(sentence) for sentence in contents_clean]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "#使用gensim中的lda模型，可以参考：https://blog.csdn.net/angela2016/article/details/78208754\n",
    "lda = gensim.models.ldamodel.LdaModel(corpus=corpus, id2word=dictionary, num_topics=20) #类似Kmeans自己指定K值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.007*\"中国\" + 0.006*\"文化\" + 0.006*\"中\" + 0.004*\"Ｓ\" + 0.004*\"Ｍ\"\n"
     ]
    }
   ],
   "source": [
    "#一号分类结果（一共分了20个主题）\n",
    "#在一号中，找出前5个\n",
    "print (lda.print_topic(1, topn=5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.008*\"中\" + 0.004*\"爆料\" + 0.004*\"卫视\" + 0.003*\"电影\" + 0.003*\"中国\"\n",
      "0.007*\"中国\" + 0.006*\"文化\" + 0.006*\"中\" + 0.004*\"Ｓ\" + 0.004*\"Ｍ\"\n",
      "0.006*\"中\" + 0.003*\"说\" + 0.003*\"恋情\" + 0.003*\"导演\" + 0.003*\"评委\"\n",
      "0.005*\"中\" + 0.003*\"公司\" + 0.003*\"中国\" + 0.003*\"Ｔ\" + 0.002*\"航母\"\n",
      "0.006*\"张绍\" + 0.004*\"乳房\" + 0.003*\"高考\" + 0.003*\"中\" + 0.003*\"Ｖ\"\n",
      "0.004*\"中\" + 0.003*\"中国\" + 0.002*\"工作\" + 0.002*\"恋情\" + 0.002*\"比赛\"\n",
      "0.030*\"ａ\" + 0.030*\"ｅ\" + 0.021*\"ｏ\" + 0.021*\"ｉ\" + 0.020*\"ｎ\"\n",
      "0.007*\"万\" + 0.007*\"该剧\" + 0.005*\"号\" + 0.003*\"中\" + 0.003*\"男士\"\n",
      "0.006*\"说\" + 0.005*\"中\" + 0.003*\"网友\" + 0.002*\"导演\" + 0.002*\"演员\"\n",
      "0.005*\"中\" + 0.004*\"食物\" + 0.003*\"吃\" + 0.003*\"撒\" + 0.003*\"含有\"\n",
      "0.007*\"男人\" + 0.005*\"孩子\" + 0.005*\"说\" + 0.004*\"爱\" + 0.004*\"女人\"\n",
      "0.008*\"中国\" + 0.005*\"中\" + 0.003*\"说\" + 0.003*\"Ｓ\" + 0.003*\"Ｄ\"\n",
      "0.014*\"电影\" + 0.006*\"导演\" + 0.006*\"中\" + 0.006*\"影片\" + 0.004*\"票房\"\n",
      "0.008*\"中\" + 0.007*\"节目\" + 0.005*\"中国\" + 0.003*\"作品\" + 0.002*\"主持人\"\n",
      "0.012*\"考生\" + 0.004*\"学生\" + 0.004*\"中\" + 0.003*\"官兵\" + 0.003*\"Ｍ\"\n",
      "0.006*\"中\" + 0.005*\"比赛\" + 0.004*\"戏\" + 0.003*\"说\" + 0.003*\"拍摄\"\n",
      "0.004*\"说\" + 0.004*\"学生\" + 0.004*\"中国\" + 0.004*\"比赛\" + 0.004*\"中\"\n",
      "0.007*\"男人\" + 0.007*\"中\" + 0.006*\"女人\" + 0.006*\"观众\" + 0.005*\"说\"\n",
      "0.006*\"肌肤\" + 0.005*\"皮肤\" + 0.005*\"Ｓ\" + 0.004*\"电视剧\" + 0.004*\"中\"\n",
      "0.004*\"中\" + 0.004*\"说\" + 0.004*\"孩子\" + 0.003*\"志愿\" + 0.003*\"工作\"\n"
     ]
    }
   ],
   "source": [
    "# 20个主题中，排名前5个主题词\n",
    "for topic in lda.print_topics(num_topics=20, num_words=5):\n",
    "    print (topic[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 10）贝叶斯分类器对新闻数据进行分类"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>contents_clean</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[经销商, 电话, 试驾, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, 路, 号, 转, ...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 二月, 公布, 最新, Ｍ, Ｉ, Ｎ, Ｉ, 新, 概念...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[清仓, 甩卖, 一汽, 夏利, Ｎ, 威志, Ｖ, 低至, 万, 启新, 中国, 一汽, ...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[日内瓦, 车展, 见到, 高尔夫, 家族, 新, 成员, 高尔夫, 敞篷版, 款, 全新,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>[经销商, 电话, 试驾, 订车, 仕, 嗍, √, 焖, 市, 秦城区, 岷山, 路, 号...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>[经销商, 电话, 试驾, 订车, 愣省, 东莞市, 南, 城区, 莞太, 大道, 白马, ...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>[自主, 品牌, 崛起, 是从, 模仿, 世界, 老牌, 名车, 一件, 坏事, 这是, 一...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>[华泰, 推出, 自主, 轿车, Ｂ, 一款, 自主, Ｓ, Ｕ, Ｖ, 宝利格, 下旬, ...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>[爱美, 之心, 人皆有之, 汽车, 外观, 炫感度, 冲击力, 第一眼, 接触, 消费者,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>[Ｍ, ａ, ｚ, ｄ, ａ, Ｍ, Ｔ, 豪华型, 款, 最新, 报价]</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>[西雅特, Ｌ, Ｅ, Ｏ, Ｎ, 上市, 喜欢, 大众, 喜欢, 运动, 朋友, 选择, ...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>[网上, 车市, 九江, 站, 讯, 理念, Ｓ, 广汽, 本田, 自主, 品牌, 首款, ...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>[经销商, 电话, 试驾, 订车, 啥际, 谢鸪, 险疚, 髀, 罚, 福叮岛, 牛, 矗,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>[Ｇ, ａ, ｌ, ｕ, ｅ, ｃ, ｏ, ｎ, ｖ, ｅ, ｒ, ｔ, ｉ, ｂ, ｌ, ｅ]</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>[业务范围, 新车, Ｊ, 谌, ㄆ, 放疲骸, 行, 刂, 罚, 骸, ∥, 诶, 铺厥,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>[经销商, 电话, 试驾, 订车, 倍, 环, 西路, 号, 中华, 大街, 北二环, 西行...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>[网上, 车市, 深圳, 站, 讯, 韩国, 进口, 双龙, 汽车, 火热, 销售, 主打,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>[雷诺, 款, Ｅ, ｓ, ｃ, ａ, ｐ, ｅ, Ｍ, Ｐ, Ｖ, 曝光, 新款, 外观,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>[克莱斯勒, 车型, 海外, 上市, 近日, 官方, 公布, 该车, 敞篷版, 细节, 图片...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>[两年, 一度, 北京, 国际, 车展, 今日, 正式, 拉开, 帷幕, 之际, 提前, 带...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>[上市, 时间, Ｗ, 刺, 在售, Ｅ, 帕浚, 海埃, Ｌ, 敌统, 叽纾撼, ぃ, 罚...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>[关注, 雪铁龙, Ｃ, 用户, 关注]</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>[喜欢, 美式, 风格, 汽车, 朋友, 美国, 车, 欧洲, 化, 很难, 新车, 辈出,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>[奥迪, Ｓ, ｐ, ｏ, ｒ, ｔ, ｂ, ａ, ｃ, ｋ, 集轿, 跑车, 运动, 优雅...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>[呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>[福建省, 公务员, 局, 三菱, 君阁, 交车, 仪式, 华夏, 汽车城, 福建, 东南,...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>[Ｈ, Ｋ, 车型, 档手, 一体, Ｓ, 秃, 模海, Ｌ, Ｋ, Ｍ, Ｖ, 傅技, 郏...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>[暂时, 商旅, Ｘ, Ｍ, Ｌ, Ｅ, Ｈ, 报价, 网上, 车市, 查看, 报价]</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>[车辆, 多元化, 需求, 拥有, 舒适性, 空间, 多功能, 一身, Ｓ, Ｕ, Ｖ, 车...</td>\n",
       "      <td>汽车</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4970</th>\n",
       "      <td>[我用, 一句, 话, 搞定, 老公, 情人, 张, 瞎, 老实人, ⑸, 那件事, 告诉,...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4971</th>\n",
       "      <td>[嘉人网, 网友, 清洁, 完, 肌肤, 先, 化妆水, 肌肤, 第二层, 清洁, 肌肤, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4972</th>\n",
       "      <td>[时尚, 解析, 可爱, 娃娃, 款, 连衣裙, 搭配, 复古, 背包, 感觉, 清新, 怡...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4973</th>\n",
       "      <td>[Ｌ, ｅ, ｅ, ｌ, ｅ, ｅ, Ｓ, ｏ, ｂ, ｉ, ｅ, ｓ, ｋ, ｉ, 蹋, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4974</th>\n",
       "      <td>[近日, 一名, 婴儿, 淋雨, 乞讨, 遭, 网友, 曝光, 警方, 称, 母子, 关系,...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4975</th>\n",
       "      <td>[伴随, 中国, 经济, 实力, 崛起, 中国, 奢侈品, 消费, 需求, 持续, 走强, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4976</th>\n",
       "      <td>[陈凯歌, 陈红, 圈内人, 羡慕, 甜蜜, 夫妻, 档, 结婚, 十年, 婚姻, 异常, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4977</th>\n",
       "      <td>[导语, 经济, 之声, 天下, 财经, 报道, 法国, 奢侈品, 巨头, 爱马仕, 集团,...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4978</th>\n",
       "      <td>[感冒, 打喷嚏, 吃, 蔬菜, 提高, 机体, 免疫力, 解决, 肌肤, 干燥, 腹泻, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4979</th>\n",
       "      <td>[近日, 财付, 通, Ｃ, ｙ, ｂ, ｅ, ｒ, ｓ, ｏ, ｕ, ｒ, ｃ, ｅ, ｓ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4980</th>\n",
       "      <td>[设计师, Ｒ, ａ, ｌ, ｐ, ｈ, Ｌ, ａ, ｕ, ｒ, ｅ, ｎ, 时装设计, 工...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4981</th>\n",
       "      <td>[导读, 岁, 女孩, 细瓷, 岁, 女人, 粗瓷, 岁, 女人, 塑料, 耐, 摔, 矫情...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4982</th>\n",
       "      <td>[夏天, 常喝, 绿豆汤, 清热解毒, 绿豆, 解百, 毒, 重金属, 中毒, 药物中毒, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4983</th>\n",
       "      <td>[电眼, 美女, 林熙蕾, 马尔代夫, 低调, 出嫁, 邀请, 三十多位, 亲友, 出席, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4984</th>\n",
       "      <td>[纽约, Ｇ, ｏ, ｖ, ｅ, ｒ, ｎ, ｏ, ｒ, ｓ, Ｂ, ａ, ｌ, ｌ, 音乐...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4985</th>\n",
       "      <td>[窦唯, 倪震, 周星驰, 叶倩文, 盘点, 偷情, 捉, 明星, 张, Ｍ, 醴疲涸, 撞...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4986</th>\n",
       "      <td>[女人, 保养, 意识, 中, 总, 抗老, 件, 昂贵, 难度, 高, 事情, 抗衰老, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4987</th>\n",
       "      <td>[乳房, 保养, 秘诀, 勤快, 作息, 不间断, 保养, 乳房, 下垂, 远离, 乳房, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4988</th>\n",
       "      <td>[脑是, 人体, 高级神经, 活动, 中枢, 感觉器官, 周身, 神经系统, 调节, 人体,...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4989</th>\n",
       "      <td>[导语, 最新, Ｗ, ｏ, ｒ, ｌ, ｄ, Ｗ, ａ, ｔ, ｃ, ｈ, Ｒ, ｅ, ｐ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4990</th>\n",
       "      <td>[责任编辑, 李昉, 王静]</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4991</th>\n",
       "      <td>[导读, 每月, 花, 护肤品, 开支, 几位数, 鼓囊囊, 钱包, 够买, 护肤品, 真,...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4992</th>\n",
       "      <td>[过时, 恋爱, 法则, 小心, 甩, 张, 罚, 电视, 两, 人均, 园, 敕, 值募,...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4993</th>\n",
       "      <td>[三文鱼, 去年, 世界卫生组织, 发表, 一项, 调查, 指出, 日本, 亚洲区, 中, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4994</th>\n",
       "      <td>[铅笔, 裙, 弹性, 窄, 裙, 英文名, Ｐ, ｅ, ｎ, ｃ, ｉ, ｌ, Ｓ, ｋ,...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4995</th>\n",
       "      <td>[天气, 炎热, 补水, 变得, 美国, 跑步, 世界, 杂志, 报道, 喝水, 身体, 补...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4996</th>\n",
       "      <td>[不想, 说, 话, 刺激, 说, 做, 只能, 走, 离开, 伤心地, 想起, 一句, 话...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4997</th>\n",
       "      <td>[岁, 刘晓庆, 最新, 嫩照, Ｏ, 衷, 诘, 牧跸, 庆, 看不出, 岁, 秒杀, 刘...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4998</th>\n",
       "      <td>[导语, 做, 爸爸, 一种, 幸福, 无论是, 领养, 亲生, 更何况, 影视剧, 中, ...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4999</th>\n",
       "      <td>[全球, 最美, 女人, 合成图, 国, 整形外科, 教授, 李承哲, 国际, 学术, 杂志...</td>\n",
       "      <td>时尚</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5000 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         contents_clean label\n",
       "0     [经销商, 电话, 试驾, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, 路, 号, 转, ...    汽车\n",
       "1     [呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...    汽车\n",
       "2     [Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 二月, 公布, 最新, Ｍ, Ｉ, Ｎ, Ｉ, 新, 概念...    汽车\n",
       "3     [清仓, 甩卖, 一汽, 夏利, Ｎ, 威志, Ｖ, 低至, 万, 启新, 中国, 一汽, ...    汽车\n",
       "4     [日内瓦, 车展, 见到, 高尔夫, 家族, 新, 成员, 高尔夫, 敞篷版, 款, 全新,...    汽车\n",
       "5     [经销商, 电话, 试驾, 订车, 仕, 嗍, √, 焖, 市, 秦城区, 岷山, 路, 号...    汽车\n",
       "6     [经销商, 电话, 试驾, 订车, 愣省, 东莞市, 南, 城区, 莞太, 大道, 白马, ...    汽车\n",
       "7     [自主, 品牌, 崛起, 是从, 模仿, 世界, 老牌, 名车, 一件, 坏事, 这是, 一...    汽车\n",
       "8     [华泰, 推出, 自主, 轿车, Ｂ, 一款, 自主, Ｓ, Ｕ, Ｖ, 宝利格, 下旬, ...    汽车\n",
       "9     [爱美, 之心, 人皆有之, 汽车, 外观, 炫感度, 冲击力, 第一眼, 接触, 消费者,...    汽车\n",
       "10                [Ｍ, ａ, ｚ, ｄ, ａ, Ｍ, Ｔ, 豪华型, 款, 最新, 报价]    汽车\n",
       "11    [西雅特, Ｌ, Ｅ, Ｏ, Ｎ, 上市, 喜欢, 大众, 喜欢, 运动, 朋友, 选择, ...    汽车\n",
       "12    [网上, 车市, 九江, 站, 讯, 理念, Ｓ, 广汽, 本田, 自主, 品牌, 首款, ...    汽车\n",
       "13    [经销商, 电话, 试驾, 订车, 啥际, 谢鸪, 险疚, 髀, 罚, 福叮岛, 牛, 矗,...    汽车\n",
       "14     [Ｇ, ａ, ｌ, ｕ, ｅ, ｃ, ｏ, ｎ, ｖ, ｅ, ｒ, ｔ, ｉ, ｂ, ｌ, ｅ]    汽车\n",
       "15    [业务范围, 新车, Ｊ, 谌, ㄆ, 放疲骸, 行, 刂, 罚, 骸, ∥, 诶, 铺厥,...    汽车\n",
       "16    [经销商, 电话, 试驾, 订车, 倍, 环, 西路, 号, 中华, 大街, 北二环, 西行...    汽车\n",
       "17    [网上, 车市, 深圳, 站, 讯, 韩国, 进口, 双龙, 汽车, 火热, 销售, 主打,...    汽车\n",
       "18    [雷诺, 款, Ｅ, ｓ, ｃ, ａ, ｐ, ｅ, Ｍ, Ｐ, Ｖ, 曝光, 新款, 外观,...    汽车\n",
       "19    [克莱斯勒, 车型, 海外, 上市, 近日, 官方, 公布, 该车, 敞篷版, 细节, 图片...    汽车\n",
       "20    [两年, 一度, 北京, 国际, 车展, 今日, 正式, 拉开, 帷幕, 之际, 提前, 带...    汽车\n",
       "21    [上市, 时间, Ｗ, 刺, 在售, Ｅ, 帕浚, 海埃, Ｌ, 敌统, 叽纾撼, ぃ, 罚...    汽车\n",
       "22                                 [关注, 雪铁龙, Ｃ, 用户, 关注]    汽车\n",
       "23    [喜欢, 美式, 风格, 汽车, 朋友, 美国, 车, 欧洲, 化, 很难, 新车, 辈出,...    汽车\n",
       "24    [奥迪, Ｓ, ｐ, ｏ, ｒ, ｔ, ｂ, ａ, ｃ, ｋ, 集轿, 跑车, 运动, 优雅...    汽车\n",
       "25    [呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...    汽车\n",
       "26    [福建省, 公务员, 局, 三菱, 君阁, 交车, 仪式, 华夏, 汽车城, 福建, 东南,...    汽车\n",
       "27    [Ｈ, Ｋ, 车型, 档手, 一体, Ｓ, 秃, 模海, Ｌ, Ｋ, Ｍ, Ｖ, 傅技, 郏...    汽车\n",
       "28          [暂时, 商旅, Ｘ, Ｍ, Ｌ, Ｅ, Ｈ, 报价, 网上, 车市, 查看, 报价]    汽车\n",
       "29    [车辆, 多元化, 需求, 拥有, 舒适性, 空间, 多功能, 一身, Ｓ, Ｕ, Ｖ, 车...    汽车\n",
       "...                                                 ...   ...\n",
       "4970  [我用, 一句, 话, 搞定, 老公, 情人, 张, 瞎, 老实人, ⑸, 那件事, 告诉,...    时尚\n",
       "4971  [嘉人网, 网友, 清洁, 完, 肌肤, 先, 化妆水, 肌肤, 第二层, 清洁, 肌肤, ...    时尚\n",
       "4972  [时尚, 解析, 可爱, 娃娃, 款, 连衣裙, 搭配, 复古, 背包, 感觉, 清新, 怡...    时尚\n",
       "4973  [Ｌ, ｅ, ｅ, ｌ, ｅ, ｅ, Ｓ, ｏ, ｂ, ｉ, ｅ, ｓ, ｋ, ｉ, 蹋, ...    时尚\n",
       "4974  [近日, 一名, 婴儿, 淋雨, 乞讨, 遭, 网友, 曝光, 警方, 称, 母子, 关系,...    时尚\n",
       "4975  [伴随, 中国, 经济, 实力, 崛起, 中国, 奢侈品, 消费, 需求, 持续, 走强, ...    时尚\n",
       "4976  [陈凯歌, 陈红, 圈内人, 羡慕, 甜蜜, 夫妻, 档, 结婚, 十年, 婚姻, 异常, ...    时尚\n",
       "4977  [导语, 经济, 之声, 天下, 财经, 报道, 法国, 奢侈品, 巨头, 爱马仕, 集团,...    时尚\n",
       "4978  [感冒, 打喷嚏, 吃, 蔬菜, 提高, 机体, 免疫力, 解决, 肌肤, 干燥, 腹泻, ...    时尚\n",
       "4979  [近日, 财付, 通, Ｃ, ｙ, ｂ, ｅ, ｒ, ｓ, ｏ, ｕ, ｒ, ｃ, ｅ, ｓ...    时尚\n",
       "4980  [设计师, Ｒ, ａ, ｌ, ｐ, ｈ, Ｌ, ａ, ｕ, ｒ, ｅ, ｎ, 时装设计, 工...    时尚\n",
       "4981  [导读, 岁, 女孩, 细瓷, 岁, 女人, 粗瓷, 岁, 女人, 塑料, 耐, 摔, 矫情...    时尚\n",
       "4982  [夏天, 常喝, 绿豆汤, 清热解毒, 绿豆, 解百, 毒, 重金属, 中毒, 药物中毒, ...    时尚\n",
       "4983  [电眼, 美女, 林熙蕾, 马尔代夫, 低调, 出嫁, 邀请, 三十多位, 亲友, 出席, ...    时尚\n",
       "4984  [纽约, Ｇ, ｏ, ｖ, ｅ, ｒ, ｎ, ｏ, ｒ, ｓ, Ｂ, ａ, ｌ, ｌ, 音乐...    时尚\n",
       "4985  [窦唯, 倪震, 周星驰, 叶倩文, 盘点, 偷情, 捉, 明星, 张, Ｍ, 醴疲涸, 撞...    时尚\n",
       "4986  [女人, 保养, 意识, 中, 总, 抗老, 件, 昂贵, 难度, 高, 事情, 抗衰老, ...    时尚\n",
       "4987  [乳房, 保养, 秘诀, 勤快, 作息, 不间断, 保养, 乳房, 下垂, 远离, 乳房, ...    时尚\n",
       "4988  [脑是, 人体, 高级神经, 活动, 中枢, 感觉器官, 周身, 神经系统, 调节, 人体,...    时尚\n",
       "4989  [导语, 最新, Ｗ, ｏ, ｒ, ｌ, ｄ, Ｗ, ａ, ｔ, ｃ, ｈ, Ｒ, ｅ, ｐ...    时尚\n",
       "4990                                     [责任编辑, 李昉, 王静]    时尚\n",
       "4991  [导读, 每月, 花, 护肤品, 开支, 几位数, 鼓囊囊, 钱包, 够买, 护肤品, 真,...    时尚\n",
       "4992  [过时, 恋爱, 法则, 小心, 甩, 张, 罚, 电视, 两, 人均, 园, 敕, 值募,...    时尚\n",
       "4993  [三文鱼, 去年, 世界卫生组织, 发表, 一项, 调查, 指出, 日本, 亚洲区, 中, ...    时尚\n",
       "4994  [铅笔, 裙, 弹性, 窄, 裙, 英文名, Ｐ, ｅ, ｎ, ｃ, ｉ, ｌ, Ｓ, ｋ,...    时尚\n",
       "4995  [天气, 炎热, 补水, 变得, 美国, 跑步, 世界, 杂志, 报道, 喝水, 身体, 补...    时尚\n",
       "4996  [不想, 说, 话, 刺激, 说, 做, 只能, 走, 离开, 伤心地, 想起, 一句, 话...    时尚\n",
       "4997  [岁, 刘晓庆, 最新, 嫩照, Ｏ, 衷, 诘, 牧跸, 庆, 看不出, 岁, 秒杀, 刘...    时尚\n",
       "4998  [导语, 做, 爸爸, 一种, 幸福, 无论是, 领养, 亲生, 更何况, 影视剧, 中, ...    时尚\n",
       "4999  [全球, 最美, 女人, 合成图, 国, 整形外科, 教授, 李承哲, 国际, 学术, 杂志...    时尚\n",
       "\n",
       "[5000 rows x 2 columns]"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#课外学习“if __name__ == '__main__' ”，参考：http://blog.konghy.cn/2017/04/24/python-entry-program/\n",
    "# 1）拿到数据，导入为DataFrame格式，带标签的\n",
    "df_train=pd.DataFrame({'contents_clean':contents_clean,'label':df_news['category']})\n",
    "df_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['汽车', '财经', '科技', '健康', '体育', '教育', '文化', '军事', '娱乐', '时尚'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#2）找出多少种label\n",
    "df_train.label.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>contents_clean</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[经销商, 电话, 试驾, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, 路, 号, 转, ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 二月, 公布, 最新, Ｍ, Ｉ, Ｎ, Ｉ, 新, 概念...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[清仓, 甩卖, 一汽, 夏利, Ｎ, 威志, Ｖ, 低至, 万, 启新, 中国, 一汽, ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[日内瓦, 车展, 见到, 高尔夫, 家族, 新, 成员, 高尔夫, 敞篷版, 款, 全新,...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                      contents_clean  label\n",
       "0  [经销商, 电话, 试驾, 订车, Ｕ, 憬, 杭州, 滨江区, 江陵, 路, 号, 转, ...      1\n",
       "1  [呼叫, 热线, 服务, 邮箱, ｋ, ｆ, ｐ, ｅ, ｏ, ｐ, ｌ, ｅ, ｄ, ａ,...      1\n",
       "2  [Ｍ, Ｉ, Ｎ, Ｉ, 品牌, 二月, 公布, 最新, Ｍ, Ｉ, Ｎ, Ｉ, 新, 概念...      1\n",
       "3  [清仓, 甩卖, 一汽, 夏利, Ｎ, 威志, Ｖ, 低至, 万, 启新, 中国, 一汽, ...      1\n",
       "4  [日内瓦, 车展, 见到, 高尔夫, 家族, 新, 成员, 高尔夫, 敞篷版, 款, 全新,...      1"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#3）因为sklearn不认识这些label，需要把这些label转换为数字，用字典来做映射\n",
    "label_mapping = {\"汽车\": 1, \"财经\": 2, \"科技\": 3, \"健康\": 4, \"体育\":5, \"教育\": 6,\"文化\": 7,\"军事\": 8,\"娱乐\": 9,\"时尚\": 0}\n",
    "#把label进行替换 map()\n",
    "df_train['label'] = df_train['label'].map(label_mapping)\n",
    "df_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "#4）sklearn, 先进行数据切分，分为训练集与测试集\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(df_train['contents_clean'].values, df_train['label'].values, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'上海'"
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#x_train = x_train.flatten()，\n",
    "x_train[0][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'中新网 上海 日电 于俊 父亲节 网络 吃 一顿 电影 快餐 微 电影 爸 对不起 我爱你 定于 本月 父亲节 当天 各大 视频 网站 首映 葜 谱 鞣 剑 保慈 障蚣 钦 呓 樯 埽 ⒌ 缬 埃 ǎ 停 椋 悖 颍 铩 妫 椋 恚 称 微型 电影 新 媒体 平台 播放 状态 短时 休闲 状态 观看 完整 策划 系统 制作 体系 支持 显示 较完整 故事情节 电影 微 超短 放映 微 周期 制作 天 数周 微 规模 投资 人民币 几千 数万元 每部 内容 融合 幽默 搞怪 时尚 潮流 人文 言情 公益 教育 商业 定制 主题 单独 成篇 系列 成剧 唇 开播 微 电影 爸 对不起 我爱你 讲述 一对 父子 观念 缺少 沟通 导致 关系 父亲 传统 固执 钟情 传统 生活 方式 儿子 新派 音乐 达 习惯 晚出 早 生活 性格 张扬 叛逆 两种 截然不同 生活 方式 理念 差异 一场 父子 间 拉开序幕 子 失手 打破 父亲 心爱 物品 父亲 赶出 家门 剧情 演绎 父亲节 妹妹 哥哥 化解 父亲 这场 矛盾 映逋坏 嚼 斫 狻 ⒍ 粤 ⒌ 桨容 争执 退让 传统 尴尬 父子 尴尬 情 男人 表达 心中 那份 感恩 一杯 滤挂 咖啡 父亲节 变得 温馨 镁 缬 缮 虾 Ｎ 逄 煳 幕 传播 迪欧 咖啡 联合 出品 出品人 希望 观摩 扪心自问 父亲节 父亲 记得 父亲 生日 哪一天 父亲 爱喝 跨出 家门 那一刻 感觉 一颗 颤动 心 操劳 天下 儿女 父亲节 大声 喊出 父亲 家人 爱 完'"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#5）把训练集中的list转换为string，注意输入格式，后续需要把每条content转换为对应的向量，通过sklearn中的向量构造器\n",
    "words = []\n",
    "for line_index in range(len(x_train)):\n",
    "    try:\n",
    "        #x_train[line_index][word_index] = str(x_train[line_index][word_index])\n",
    "        #需要把list转换为string格式，可以用.join形式组合，并且用‘ ’空格取分开\n",
    "        #join()用法参考：https://blog.csdn.net/weixin_40475396/article/details/78227747\n",
    "        words.append(' '.join(x_train[line_index])) #words里面是训练数据\n",
    "    except:\n",
    "        print (line_index,word_index)\n",
    "words[0]        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3750\n"
     ]
    }
   ],
   "source": [
    "print (len(words))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['bird', 'cat', 'dog', 'fish']\n",
      "[[0 1 1 1]\n",
      " [0 2 1 0]\n",
      " [1 0 0 1]\n",
      " [1 0 0 0]]\n",
      "[2 3 2 2]\n"
     ]
    }
   ],
   "source": [
    "#先举例\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "#我有4篇文章\n",
    "#【格式要求】：这里输入为list，每篇文章用逗号分开，每篇文章里面用空格分开，不能写成字符串\n",
    "texts=[\"dog cat fish\",\"dog cat cat\",\"fish bird\", 'bird'] \n",
    "#向量构造器，实例化一个对象\n",
    "cv = CountVectorizer()\n",
    "#进行向量转换\n",
    "cv_fit=cv.fit_transform(texts)\n",
    "\n",
    "#现在语料库中不重复的词有几个，一共是4中\n",
    "print(cv.get_feature_names())\n",
    "#打印转换好的向量，解释第一行：‘bird’单词在第一篇文章中出现0次，所以第一个值为0，‘cat’出现一次，所以为1，以此类推\n",
    "print(cv_fit.toarray())\n",
    "\n",
    "\n",
    "print(cv_fit.toarray().sum(axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['bird', 'cat', 'cat cat', 'cat fish', 'dog', 'dog cat', 'dog cat cat', 'dog cat fish', 'fish', 'fish bird']\n",
      "[[0 1 0 1 1 1 0 1 1 0]\n",
      " [0 2 1 0 1 1 1 0 0 0]\n",
      " [1 0 0 0 0 0 0 0 1 1]\n",
      " [1 0 0 0 0 0 0 0 0 0]]\n",
      "[2 3 1 1 2 2 1 1 2 1]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "texts=[\"dog cat fish\",\"dog cat cat\",\"fish bird\", 'bird']\n",
    "#注意这里有 ngram_range ，可以让词组合，让种类更多更复杂，向量从原来的4维转为了9维，当然也别太多\n",
    "#通常为ngram 为 2就行\n",
    "cv = CountVectorizer(ngram_range=(1,4))\n",
    "cv_fit=cv.fit_transform(texts)\n",
    "\n",
    "print(cv.get_feature_names())\n",
    "print(cv_fit.toarray())\n",
    "\n",
    "\n",
    "print(cv_fit.toarray().sum(axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CountVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
       "        lowercase=False, max_df=1.0, max_features=4000, min_df=1,\n",
       "        ngram_range=(1, 1), preprocessor=None, stop_words=None,\n",
       "        strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
       "        tokenizer=None, vocabulary=None)"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 6）构造向量，有了上面的words的标准格式的内容，开始构建特征向量\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "vec = CountVectorizer(analyzer='word', max_features=4000,  lowercase = False)\n",
    "#fit一下\n",
    "vec.fit(words)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 7） 导入在sklearn中，把beyes拿出来，有输入特征，还有label值\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "#实例化分类器对象\n",
    "classifier = MultinomialNB()\n",
    "#把words向量传入classifier\n",
    "classifier.fit(vec.transform(words), y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'国家 公务员 考试 申论 应用文 类 试题 实质 一道 集 概括 分析 提出 解决问题 一体 综合性 试题 说 一道 客观 凝练 申发 论述 文章 题目 分析 历年 国考 申论 真题 公文 类 试题 类型 多样 包括 公文 类 事务性 文书 类 题材 从题 干 作答 材料 内容 整合 分析 无需 太 创造性 发挥 纵观 历年 申论 真题 作答 应用文 类 试题 文种 格式 作出 特别 重在 内容 考查 行文 格式 考生 平常心 面对 应用文 类 试题 准确 把握 作答 领会 内在 含义 把握 题材 主旨 材料 结构 轻松 应对 应用文 类 试题 Ｒ 弧 ⒆ 钒 盐 展文 写作 原则 Ｔ 材料 中来 应用文 类 试题 材料 总体 把握 客观 考生 材料 中来 材料 中 把握 材料 准确 理解 题材 主旨 Ｔ 政府 角度 作答 应用文 类 试题 更应 注重 政府 角度 观点 政府 角度 出发 原则 表述 观点 提出 解决 之策 考生 作答 站 政府 人员 角度 看待 提出 解决问题 Ｔ 文体 结构 形式 考查 重点 文体 结构 大部分 评分 关键点 解答 方法 薄 ⒆ ス 丶 词 明 方向 作答 题目 题干 作答 作答 方向 作答 角度 关键 向导 考生 仔细阅读 题干 作答 抓住 关键词 作答 方向 相关 要点 整理 作答 思路 年国考 地市级 真 题为 例 潦惺姓 府 宣传 推进 近海 水域 污染 整治 工作 请 给定 资料 市政府 工作人员 身份 草拟 一份 宣传 纲要 Ｒ 求 保对 宣传 内容 要点 提纲挈领 陈述 玻 体现 政府 精神 全市 各界 关心 支持 污染 整治 工作 通俗易懂 超过 字 肮 丶 词 近海 水域 污染 整治 工作 市政府 工作人员 身份 宣传 纲要 提纲挈领 陈述 体现 政府 精神 全市 各界 关心 支持 污染 整治 工作 通俗易懂 提示 归结 作答 要点 包括 污染 情况 原因 解决 对策 作答 思路 情况 原因 对策 意义 逻辑 顺序 安排 文章 结构 病 ⒋ 缶殖 龇 ⅲ 明 结构 解答 应用文 类 试题 考生 材料 整体 出发 大局 出发 高屋建瓴 把握 材料 主题 思想 事件 起因 解决 对策 阅读文章 构建 文章 结构 直至 快速 解答 场 ⒗ 硭 乘悸 罚明 逻辑 应用文 类 试题 严密 逻辑思维 情况 原因 对策 意义 考生 作答 先 弄清楚 解答 思路 统筹安排 脉络 清晰 逻辑 表达 内容 表述 础 把握 明 详略 考生 仔细阅读 分析 揣摩 应用文 类 试题 内容 答题 时要 详略 得当 主次 分明 安排 内容 增加 文章 层次感 阅卷 老师 阅卷 时能 明白 清晰 一目了然 玻埃 保蹦旯 考 考试 申论 试卷 分为 省级 地市级 两套 试卷 能力 大有 省级 申论 试题 考生 宏观 角度看 注重 深度 广度 考生 深谋远虑 地市级 试题 考生 微观 视角 观察 侧重 考查 解决 能力 考生 贯彻执行 作答 区别对待'"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 8）测试集同样的3步处理操作：内容转换为string，空格分开 --- 把内容转换为特征向量 ---- 导入贝叶斯模型传入（输入特征向量与label值）\n",
    "test_words = []\n",
    "for line_index in range(len(x_test)):\n",
    "    try:\n",
    "        #x_train[line_index][word_index] = str(x_train[line_index][word_index])\n",
    "        test_words.append(' '.join(x_test[line_index]))\n",
    "    except:\n",
    "         print (line_index,word_index)\n",
    "test_words[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.804"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#9) 基于词频向量的，进行贝叶斯，结果\n",
    "classifier.score(vec.transform(test_words), y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
       "        lowercase=False, max_df=1.0, max_features=4000, min_df=1,\n",
       "        ngram_range=(1, 1), norm='l2', preprocessor=None, smooth_idf=True,\n",
       "        stop_words=None, strip_accents=None, sublinear_tf=False,\n",
       "        token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b', tokenizer=None, use_idf=True,\n",
       "        vocabulary=None)"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 10) 另一种构造向量的方式，不采用词频，采用TF-IDF模式构造向量，发现最后结果稍好一些\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "vectorizer = TfidfVectorizer(analyzer='word', max_features=4000,  lowercase = False)\n",
    "vectorizer.fit(words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.naive_bayes import MultinomialNB\n",
    "classifier = MultinomialNB()\n",
    "classifier.fit(vectorizer.transform(words), y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8152"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#11）执行当前的结果，贝叶斯分类器的精度\n",
    "classifier.score(vectorizer.transform(test_words), y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}