word2vec_繁_full.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# coding=utf-8\n",
    "import jieba\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.decomposition import PCA\n",
    "from hanziconv import HanziConv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# file path\n",
    "filePath='/Users/youngmihuang/Downloads/corpus.txt'\n",
    "fileSegWordDonePath ='corpusSegDone.txt'\n",
    "\n",
    "# read the file by file\n",
    "fileTrainRead = []\n",
    "with open(filePath) as fileTrainRaw:\n",
    "    for line in fileTrainRaw:\n",
    "        fileTrainRead.append(HanziConv.toTraditional(line))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "50000\n",
      "100000\n",
      "150000\n",
      "200000\n",
      "250000\n",
      "300000\n",
      "350000\n",
      "400000\n",
      "450000\n",
      "500000\n",
      "550000\n",
      "600000\n",
      "650000\n",
      "700000\n",
      "750000\n",
      "800000\n",
      "850000\n",
      "900000\n",
      "950000\n",
      "1000000\n",
      "1050000\n",
      "1100000\n",
      "1150000\n",
      "1200000\n",
      "1250000\n",
      "1300000\n",
      "1350000\n",
      "1400000\n"
     ]
    }
   ],
   "source": [
    "# define this function to print a list with Chinese\n",
    "def PrintListChinese(list):\n",
    "    for i in range(len(list)):\n",
    "        print list[i],\n",
    "\n",
    "# segment word with jieba\n",
    "fileTrainSeg=[]\n",
    "for i in range(len(fileTrainRead)):\n",
    "    fileTrainSeg.append([' '.join(list(jieba.cut(fileTrainRead[i][9:-11],cut_all=False)))])\n",
    "    if i % 50000 == 0 :\n",
    "        print i"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# save the result\n",
    "with open(fileSegWordDonePath,'wb') as fW:\n",
    "    for i in range(len(fileTrainSeg)):\n",
    "        fW.write(fileTrainSeg[i][0].encode('utf-8'))\n",
    "        fW.write('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting training using file corpusSegDone.txt\n",
      "Vocab size: 842956\n",
      "Words in train file: 407852192\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "IOPub data rate exceeded.\n",
      "The notebook server will temporarily stop sending output\n",
      "to the client in order to avoid crashing it.\n",
      "To change this limit, set the config variable\n",
      "`--NotebookApp.iopub_data_rate_limit`.\n"
     ]
    }
   ],
   "source": [
    "# jieba seg words transform to word2vec\n",
    "import word2vec\n",
    "word2vec.word2vec('corpusSegDone.txt', 'corpusWord2Vec.bin', size=300,verbose=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compare to 簡體結果\n",
    "- Starting training using file corpusSegDone.txt\n",
    "- Vocab size: 507882\n",
    "- Words in train file: 391829209\n",
    "\n",
    "### 以分詞效果看來，簡體分得比較聚合；繁體相對來講有些字他無法合併在一起看，導致字詞數較多\n",
    "\n",
    "- e.g. \"晚自習\" 在簡體會被分成 \"晚自習\"；但在繁體會被分成 \"晚\" 和 \"自習\" 兩個單詞"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "２ ０ １ ２ 年 東風 標 緻 小 桔 燈 鄉 村小 學圖 書館 計劃 於 ６ 月 ２ ３ 日 － ２ ９ 日 在 湖北省 武漢市 新洲 區鳳凰鎮 郭崗小學 舉行 ， 來 自全國 各地 的 ７ 名誌 願者 帶 著 東風 標 緻 ５ ０ ８ 汽車 捐贈 的 ２ ０ ０ ０ 本新書 ， 建立 起有 一所 “ 小桔 燈鄉 村小 學圖 書館 ” 。 Ｎ 浜 菏行 輪 耷不僅 是 革命 老區 ， 也 是 全國 有名 的 建築 之 鄉 ； 這 裏 的 人 大部分 都 外 齣 從 事 建築業 ， 導 緻 這 裏 的 留守 兒童 比例 其高 。 郭崗 小學現 有 學生 １ ３ ６ 名 （ 含教學點 一處 ） ， 其中 住校 學生 ９ ８ 名 ， 全校 ９ ０ ％ 左右 的 學生 都 是 留守 兒童 。 Ｔ 諢 疃 舉行 之前 ， 學校 圖書室 裏 擺滿 瞭陳舊 的 教科 書 ， 這 裏 的 學生 在 課外 時間 沒有 課外 書 可以 閱讀 ； 在 誌 願者 問學生 喜歡 讀什麼 書 的 時候 ， 大部分 孩子 的 迴 答 是 “ 童話 書 ” ， 當再問 他們 還喜歡 什麼 圖書 的 時候 ， 孩子 迷茫 的 眼神 讓 我們 明白 由 於 沒 有 接觸 過課 外書 ， 很多 孩子 不 知道 除 瞭童 話 書 以外 還有 什麼類 彆 的 書 可以 閱讀 。 Ｎ 瞭改變 這一現 狀 ， 此次 活動 捐贈 的 圖書涵 蓋 瞭童 話 、 中外文 學 、 傳統 文學 、 注音版 和 繪畫 闆 少兒 讀物 等 ； 學校 郭校 長 也 承諾會 在 每天 的 晚 自習 時 對 全校 學生 開放 閱讀 和 藉 書 。 ４ 舜位 疃 也 得到 瞭 文化 、 經濟 各界 名人 的 支持 ， 包括 陳 丹青 、 鄧康延 、 洪晃 、 陸琪 、 蔣 方舟 、 饒雪漫 、 瀋 顥 、 瀋 南 鵬 、 田 立新 、 熊焰 、 張瓊 ， 都 為 鄉村 兒童 推薦 瞭 他們 認為 應該 閱讀 的 圖書 ； 我們 也 在 東風 標 緻 ５ ０ ８ 汽車 的 支持 下 ， 在 捐贈 的 ２ ０ ０ ０ 本書 之外 購買 瞭 上述 人士 推薦 的 圖書 ， 作為 額外 的 補充 捐贈 給學校 。 Ｔ 諞 恢 芑 疃 的 結束 時 ， 誌 願者 們 為 孩子 們 安排 瞭 彆 開生 麵 的 錶 演 ， 幫助 孩子 們 展示 一周 以來 通過 閱讀 圖書 所 編排 的 各類 節目 ； 留守 在 村 裏 的 老人 成 瞭 孩子 們 的 觀眾 ； 而 很多 父母 都 在外 打工 的 學生 ， 則在 活動後由 於 遺憾 自己 的 錶 演 沒 有 親人觀 看 而 神傷 。 對 這些 孩子 ， 誌 願者 們 拉著 他們 做 遊戲 、 聊天 ， 幫助 他們 理解 父母 的 難處 ， 鼓勵 他們 好好 學習 ， 在 即 將到 來 的 暑假 帶著 優秀 的 成績 單去 城市 和 他們 的 父母 團聚 。\n"
     ]
    }
   ],
   "source": [
    "# define this function to print a list with Chinese\n",
    "def PrintListChinese(list):\n",
    "    for i in range(len(list)):\n",
    "        print list[i],\n",
    "# to test the segment result\n",
    "PrintListChinese(fileTrainSeg[10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.08015626  0.08850129 -0.07670335 ..., -0.02626957 -0.03316621\n",
      "   0.0614953 ]\n",
      " [-0.02138876 -0.06264357 -0.09334067 ...,  0.00015479 -0.05851945\n",
      "   0.04908165]\n",
      " [-0.00995629 -0.01481206 -0.09394402 ..., -0.01619852 -0.06687199\n",
      "   0.02453057]\n",
      " ..., \n",
      " [ 0.04989554  0.08304838 -0.05676759 ..., -0.00947851  0.07963122\n",
      "  -0.03377864]\n",
      " [-0.01265288  0.07221718  0.10141564 ...,  0.01648477  0.1009445\n",
      "  -0.07451316]\n",
      " [ 0.03596683 -0.08654522  0.03320961 ..., -0.07038685  0.09750576\n",
      "  -0.03816517]]\n"
     ]
    }
   ],
   "source": [
    "# see the word2vector\n",
    "model = word2vec.load('corpusWord2Vec.bin')\n",
    "print model.vectors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "達到\n",
      "屏\n",
      "行為\n",
      "觀眾\n",
      "日訊\n"
     ]
    }
   ],
   "source": [
    "# see the example of vocabulary from 996 to 1000\n",
    "for i in range(995,1000):\n",
    "    print model.vocab[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "畢業生\n",
      "離校\n",
      "考入\n",
      "剛畢業\n",
      "大四\n",
      "學畢業\n",
      "考上\n",
      "開學\n",
      "結業\n",
      "放暑假\n"
     ]
    }
   ],
   "source": [
    "# check correlation of the specific word\n",
    "indexes = model.cosine(u'畢業')\n",
    "for index in indexes[0]:\n",
    "    print model.vocab[index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "寶寶的\n",
      "小寶寶\n",
      "孩子\n",
      "準媽媽\n",
      "寶寶在\n",
      "胎兒\n",
      "媽媽們\n",
      "小孩\n",
      "小孩子\n",
      "媽媽\n"
     ]
    }
   ],
   "source": [
    "indexes = model.cosine(u'寶寶')\n",
    "for index in indexes[0]:\n",
    "    print model.vocab[index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "坐車\n",
      "開車\n",
      "坐火車\n",
      "買票\n",
      "下車\n",
      "買菜\n",
      "上車\n",
      "搭車\n",
      "乘車\n",
      "騎車\n"
     ]
    }
   ],
   "source": [
    "indexes = model.cosine(u'打車')\n",
    "for index in indexes[0]:\n",
    "    print model.vocab[index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "姚晨\n",
      "林心如\n",
      "穎兒\n",
      "張智霖\n",
      "劉愷威\n",
      "馮紹峰\n",
      "鄧超\n",
      "周迅\n",
      "柳岩\n",
      "李晨\n"
     ]
    }
   ],
   "source": [
    "indexes = model.cosine(u'楊冪')\n",
    "for index in indexes[0]:\n",
    "    print model.vocab[index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "網易\n",
      "新浪\n",
      "百度\n",
      "優酷\n",
      "搜狗\n",
      "凡客\n",
      "窩窩團\n",
      "電商\n",
      "愛奇藝\n",
      "百視通\n"
     ]
    }
   ],
   "source": [
    "indexes = model.cosine(u'騰訊')\n",
    "for index in indexes[0]:\n",
    "    print model.vocab[index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# load the word2vec model\n",
    "rawWordVec = model.vectors\n",
    " \n",
    "# reduce the dimension of word vector\n",
    "X_reduced = PCA(n_components=2).fit_transform(rawWordVec)\n",
    " \n",
    "# show some word(center word) and it's similar words\n",
    "index1,metrics1 = model.cosine(u'畢業')\n",
    "index2,metrics2 = model.cosine(u'寶寶')\n",
    "index3,metrics3 = model.cosine(u'打車')\n",
    "index4,metrics4 = model.cosine(u'楊冪')\n",
    "index5,metrics5 = model.cosine(u'騰訊')\n",
    " \n",
    "# add the index of center word \n",
    "index01=np.where(model.vocab==u'畢業')\n",
    "index02=np.where(model.vocab==u'寶寶')\n",
    "index03=np.where(model.vocab==u'打車')\n",
    "index04=np.where(model.vocab==u'楊冪')\n",
    "index05=np.where(model.vocab==u'騰訊')\n",
    " \n",
    "index1=np.append(index1,index01)\n",
    "index2=np.append(index2,index02)\n",
    "index3=np.append(index3,index03)\n",
    "index4=np.append(index4,index04)\n",
    "index5=np.append(index5,index05)\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAD8CAYAAACfF6SlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x25ff60e90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x18fd93110>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot the result\n",
    "zhfont = matplotlib.font_manager.FontProperties(fname='/Users/youngmihuang/Downloads/wqy-microhei.ttc')\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    " \n",
    "for i in index1:\n",
    "    ax.text(X_reduced[i][0],X_reduced[i][1], model.vocab[i], fontproperties=zhfont,color='C3')\n",
    "\n",
    "for i in index2:\n",
    "    ax.text(X_reduced[i][0],X_reduced[i][1], model.vocab[i], fontproperties=zhfont,color='C1')\n",
    "\n",
    "# for i in index3:\n",
    "#     ax.text(X_reduced[i][0],X_reduced[i][1], model.vocab[i], fontproperties=zhfont,color='C7')\n",
    "\n",
    "for i in index4:\n",
    "    ax.text(X_reduced[i][0],X_reduced[i][1], model.vocab[i], fontproperties=zhfont,color='C0')\n",
    "\n",
    "# for i in index5:\n",
    "#     ax.text(X_reduced[i][0],X_reduced[i][1], model.vocab[i], fontproperties=zhfont,color='C4')\n",
    "\n",
    "ax.axis([0.10,0.43, -0.20,0.60])\n",
    "plt.figure(figsize=(60,60))\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}