20151030Demo.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "vectorizer = CountVectorizer(min_df=1)\n",
    "content = [\"How to format my hard disk\", \" Hard disk format problems \"]\n",
    "X = vectorizer.fit_transform(content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  (0, 3)\t1\n",
      "  (0, 6)\t1\n",
      "  (0, 1)\t1\n",
      "  (0, 4)\t1\n",
      "  (0, 2)\t1\n",
      "  (0, 0)\t1\n",
      "  (1, 1)\t1\n",
      "  (1, 2)\t1\n",
      "  (1, 0)\t1\n",
      "  (1, 5)\t1\n",
      "[[1 1 1 1 1 0 1]\n",
      " [1 1 1 0 0 1 0]]\n"
     ]
    }
   ],
   "source": [
    "print X\n",
    "print X.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'disk', u'format', u'hard', u'how', u'my', u'problems', u'to']\n",
      "[[1 1 1 1 1 0 1]\n",
      " [1 1 1 0 0 1 0]]\n"
     ]
    }
   ],
   "source": [
    "content = [\"How to format my hard disk\", \" Hard disk format problems \"]\n",
    "print vectorizer.get_feature_names()\n",
    "print X.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1 1]\n",
      " [1 1]\n",
      " [1 1]\n",
      " [1 0]\n",
      " [1 0]\n",
      " [0 1]\n",
      " [1 0]]\n"
     ]
    }
   ],
   "source": [
    "print(X.toarray().transpose())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['01.txt', '02.txt', '03.txt', '04.txt', '05.txt']\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "print os.listdir('toy')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "This is a toy post about machine learning. Actually, it contains not much interesting stuff.\n"
     ]
    }
   ],
   "source": [
    "f = open('toy/01.txt', 'r')\n",
    "print f.read()\n",
    "f.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['This is a toy post about machine learning. Actually, it contains not much interesting stuff.', 'Imaging databases provide storage capabilities.', 'Most imaging databases safe images permanently.', 'Imaging databases store data.', 'Imaging databases store data. Imaging databases store data. Imaging databases store data.']\n"
     ]
    }
   ],
   "source": [
    "DIR = 'toy'\n",
    "posts = [] \n",
    "for f in os.listdir('toy'):\n",
    "    #print os.path.join(DIR, f)\n",
    "    posts.append(open(os.path.join(DIR, f)).read())\n",
    "print posts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "posts = [open(os.path.join(DIR, f)).read() for f in os.listdir('toy')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "vectorizer = CountVectorizer(min_df=1)\n",
    "\n",
    "X_train = vectorizer.fit_transform(posts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'about', u'actually', u'capabilities', u'contains', u'data', u'databases', u'images', u'imaging', u'interesting', u'is', u'it', u'learning', u'machine', u'most', u'much', u'not', u'permanently', u'post', u'provide', u'safe', u'storage', u'store', u'stuff', u'this', u'toy']\n"
     ]
    }
   ],
   "source": [
    "print vectorizer.get_feature_names()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5, 25)\n"
     ]
    }
   ],
   "source": [
    "print X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'about', u'actually', u'capabilities', u'contains', u'data', u'databases', u'images', u'imaging', u'interesting', u'is', u'it', u'learning', u'machine', u'most', u'much', u'not', u'permanently', u'post', u'provide', u'safe', u'storage', u'store', u'stuff', u'this', u'toy']\n",
      "[[1 1 0 1 0 0 0 0 1 1 1 1 1 0 1 1 0 1 0 0 0 0 1 1 1]\n",
      " [0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0]\n",
      " [0 0 0 0 0 1 1 1 0 0 0 0 0 1 0 0 1 0 0 1 0 0 0 0 0]\n",
      " [0 0 0 0 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0]\n",
      " [0 0 0 0 3 3 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0]]\n"
     ]
    }
   ],
   "source": [
    "print vectorizer.get_feature_names()\n",
    "print X_train.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "2\n"
     ]
    }
   ],
   "source": [
    "a,b = 1,2\n",
    "print a\n",
    "print b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2 1\n"
     ]
    }
   ],
   "source": [
    "a,b=1,2\n",
    "c = a\n",
    "a = b\n",
    "b = c\n",
    "print a,b\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2 1\n"
     ]
    }
   ],
   "source": [
    "a,b = b,a\n",
    "print a, b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5, 25)\n",
      "5\n",
      "25\n"
     ]
    }
   ],
   "source": [
    "print X_train.shape\n",
    "#num_samples, num_features = X_train.shape\n",
    "num_samples, num_features = X_train.shape\n",
    "print num_samples\n",
    "print num_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#samples: 5, #features: 25\n",
      "[u'about', u'actually', u'capabilities', u'contains', u'data', u'databases', u'images', u'imaging', u'interesting', u'is', u'it', u'learning', u'machine', u'most', u'much', u'not', u'permanently', u'post', u'provide', u'safe', u'storage', u'store', u'stuff', u'this', u'toy']\n"
     ]
    }
   ],
   "source": [
    "print(\"#samples: %d, #features: %d\" % (num_samples,num_features)) \n",
    "print(vectorizer.get_feature_names())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "new_post = \"imaging databases\"\n",
    "new_post_vec = vectorizer.transform([new_post])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 25)\n"
     ]
    }
   ],
   "source": [
    "print new_post_vec.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]\n"
     ]
    }
   ],
   "source": [
    "print new_post_vec.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import scipy as sp\n",
    "def dist_raw(v1, v2):\n",
    "    delta = v1-v2\n",
    "    return sp.linalg.norm(delta.toarray())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== Post 0 with dist=4.00: This is a toy post about machine learning. Actually, it contains not much interesting stuff.\n",
      "=== Post 1 with dist=1.73: Imaging databases provide storage capabilities.\n",
      "=== Post 2 with dist=2.00: Most imaging databases safe images permanently.\n",
      "=== Post 3 with dist=1.41: Imaging databases store data.\n",
      "=== Post 4 with dist=5.10: Imaging databases store data. Imaging databases store data. Imaging databases store data.\n",
      "Best post is 3 with dist=1.41\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "best_doc = None\n",
    "best_dist = sys.maxint\n",
    "best_i = None\n",
    "for i in range(0, num_samples):\n",
    "    post = posts[i]\n",
    "    if post==new_post:\n",
    "        continue\n",
    "    post_vec = X_train.getrow(i)\n",
    "    d = dist_raw(post_vec, new_post_vec)\n",
    "    print \"=== Post %i with dist=%.2f: %s\"%(i, d, post)\n",
    "    if d<best_dist:\n",
    "        best_dist = d\n",
    "        best_i = i\n",
    "print(\"Best post is %i with dist=%.2f\"%(best_i, best_dist))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def dist_norm(v1, v2):\n",
    "    v1_normalized = v1/sp.linalg.norm(v1.toarray())\n",
    "    v2_normalized = v2/sp.linalg.norm(v2.toarray())\n",
    "    delta = v1_normalized - v2_normalized\n",
    "    return sp.linalg.norm(delta.toarray())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.76536686473\n"
     ]
    }
   ],
   "source": [
    "print dist_norm(X_train[4], new_post_vec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== Post 0 with dist=1.41: This is a toy post about machine learning. Actually, it contains not much interesting stuff.\n",
      "=== Post 1 with dist=0.86: Imaging databases provide storage capabilities.\n",
      "=== Post 2 with dist=0.92: Most imaging databases safe images permanently.\n",
      "=== Post 3 with dist=0.77: Imaging databases store data.\n",
      "=== Post 4 with dist=0.77: Imaging databases store data. Imaging databases store data. Imaging databases store data.\n",
      "Best post is 3 with dist=0.77\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "best_doc = None\n",
    "best_dist = sys.maxint\n",
    "best_i = None\n",
    "for i in range(0, num_samples):\n",
    "    post = posts[i]\n",
    "    if post==new_post:\n",
    "        continue\n",
    "    post_vec = X_train.getrow(i)\n",
    "    d = dist_norm(post_vec, new_post_vec)\n",
    "    print \"=== Post %i with dist=%.2f: %s\"%(i, d, post)\n",
    "    if d<best_dist:\n",
    "        best_dist = d\n",
    "        best_i = i\n",
    "print(\"Best post is %i with dist=%.2f\"%(best_i, best_dist))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "318\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['a',\n",
       " 'about',\n",
       " 'above',\n",
       " 'across',\n",
       " 'after',\n",
       " 'afterwards',\n",
       " 'again',\n",
       " 'against',\n",
       " 'all',\n",
       " 'almost',\n",
       " 'alone',\n",
       " 'along',\n",
       " 'already',\n",
       " 'also',\n",
       " 'although',\n",
       " 'always',\n",
       " 'am',\n",
       " 'among',\n",
       " 'amongst',\n",
       " 'amoungst']"
      ]
     },
     "execution_count": 252,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vectorizer = CountVectorizer(min_df=1, stop_words='english')\n",
    "print len(vectorizer.get_stop_words())\n",
    "sorted(vectorizer.get_stop_words())[0:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'actually', u'capabilities', u'contains', u'data', u'databases', u'images', u'imaging', u'interesting', u'learning', u'machine', u'permanently', u'post', u'provide', u'safe', u'storage', u'store', u'stuff', u'toy']\n"
     ]
    }
   ],
   "source": [
    "X_train = vectorizer.fit_transform(posts)\n",
    "print vectorizer.get_feature_names()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "new_post = \"imaging databases\"\n",
    "new_post_vec = vectorizer.transform([new_post])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== Post 0 with dist=1.41: This is a toy post about machine learning. Actually, it contains not much interesting stuff.\n",
      "=== Post 1 with dist=0.86: Imaging databases provide storage capabilities.\n",
      "=== Post 2 with dist=0.86: Most imaging databases safe images permanently.\n",
      "=== Post 3 with dist=0.77: Imaging databases store data.\n",
      "=== Post 4 with dist=0.77: Imaging databases store data. Imaging databases store data. Imaging databases store data.\n",
      "Best post is 3 with dist=0.77\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "best_doc = None\n",
    "best_dist = sys.maxint\n",
    "best_i = None\n",
    "for i in range(0, num_samples):\n",
    "    post = posts[i]\n",
    "    if post==new_post:\n",
    "        continue\n",
    "    post_vec = X_train.getrow(i)\n",
    "    d = dist_norm(post_vec, new_post_vec)\n",
    "    print \"=== Post %i with dist=%.2f: %s\"%(i, d, post)\n",
    "    if d<best_dist:\n",
    "        best_dist = d\n",
    "        best_i = i\n",
    "print(\"Best post is %i with dist=%.2f\"%(best_i, best_dist))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "graphic\n",
      "imag\n",
      "imag\n",
      "imagin\n",
      "imagin\n"
     ]
    }
   ],
   "source": [
    "import nltk.stem\n",
    "s= nltk.stem.SnowballStemmer('english')\n",
    "print s.stem(\"graphics\")\n",
    "print s.stem(\"imaging\")\n",
    "print s.stem(\"image\")\n",
    "print s.stem(\"imagination\")\n",
    "print s.stem(\"imagine\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import nltk.stem\n",
    "english_stemmer = nltk.stem.SnowballStemmer('english')\n",
    "class StemmedCountVectorizer(CountVectorizer):\n",
    "    def build_analyzer(self):\n",
    "        analyzer = super(StemmedCountVectorizer, self).build_analyzer()\n",
    "        return lambda doc: (english_stemmer.stem(w) for w in analyzer(doc))\n",
    "vectorizer = StemmedCountVectorizer(min_df=1, stop_words='english')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'actual', u'capabl', u'contain', u'data', u'databas', u'imag', u'interest', u'learn', u'machin', u'perman', u'post', u'provid', u'safe', u'storag', u'store', u'stuff', u'toy']\n"
     ]
    }
   ],
   "source": [
    "X_train = vectorizer.fit_transform(posts)\n",
    "print vectorizer.get_feature_names()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import sqlite3\n",
    "db = sqlite3.connect('news2.sqlite') \n",
    "cur = db.cursor()\n",
    "cur.execute('select title, summary, category from news_entry')\n",
    "allNews = cur.fetchall()\n",
    "\n",
    "corpus = []\n",
    "tags = []\n",
    "title = []\n",
    "for rec in allNews:\n",
    "    #if (rec[2] == '娛樂'.decode('utf-8')) or \\\n",
    "    #   (rec[2] == '社會'.decode('utf-8')) or \\\n",
    "    #   (rec[2] == '財經'.decode('utf-8'))  :\n",
    "    corpus.append(' '.join(jieba.cut(rec[1])))\n",
    "    tags.append(rec[2]) \n",
    "cur.close()\n",
    "db.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "87\n"
     ]
    }
   ],
   "source": [
    "print tags.count('財經'.decode('utf-8'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 336,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "901\n"
     ]
    }
   ],
   "source": [
    "print len(corpus)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 337,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import jieba\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "vectorizer = CountVectorizer() \n",
    "X = vectorizer.fit_transform(corpus)\n",
    "word = vectorizer.get_feature_names() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 338,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "samples, features =  X.shape\n",
    "X2= X.toarray().transpose()\n",
    "#print X2\n",
    "#print X.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 339,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.cross_validation import train_test_split\n",
    "train_data, test_data, train_tag, test_tag = train_test_split(X, tags, test_size=0.50, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 340,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(450, 35298)\n",
      "(451, 35298)\n"
     ]
    }
   ],
   "source": [
    "print train_data.shape\n",
    "print test_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 341,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0 0 ..., 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "print X[0].toarray().flatten()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 342,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultinomialNB(alpha=0.01, class_prior=None, fit_prior=True)"
      ]
     },
     "execution_count": 342,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.naive_bayes import MultinomialNB\n",
    "import numpy\n",
    "\n",
    "clf = MultinomialNB(alpha=0.01)\n",
    "clf.fit(train_data,train_tag)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 343,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "pred = clf.predict(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 344,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#print test_data[2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.60088691796\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "print accuracy_score(test_tag, pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 356,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 4  2  0  0  2  0  0  0  4  0  1  1  1  1  0  0]\n",
      " [ 1  8  0  0  0  0  0  0  1  0  0  1  0  0  0  0]\n",
      " [ 0  1 31  0  0  6  1  1  9  0  3  4  2  0  1  0]\n",
      " [ 0  0  0  6  0  0  0  0  1  0  0  1  0  0  0  0]\n",
      " [ 1  1  0  0  7  0  0  1 17  0  0  1  1  0  0  0]\n",
      " [ 0  0  0  0  2  2  0  0  2  0  0  1  0  0  0  0]\n",
      " [ 0  0  0  0  0  0 34  0  5  0  1  1  2  1  0  0]\n",
      " [ 0  0  0  0  0  0  0 12  1  0  0  1  0  0  0  0]\n",
      " [ 0  1  0  0  0  0  0  0  3  0  0  0  0  0  0  0]\n",
      " [ 0  0  0  0  0  0  0  1  0  0  0  0  0  0  0  0]\n",
      " [ 3  2  1  4  1  2  4  1 22  0 38  5  4  3  1  0]\n",
      " [ 0  0  0  0  0  0  0  0  1  0  0  8  0  0  0  0]\n",
      " [ 0  0  0  0  0  3  3  0 12  0  1  1 53  0  0  0]\n",
      " [ 0  0  0  0  0  0  0  0  3  0  1  1  0  9  1  1]\n",
      " [ 0  0  1  0  0  1  0  0  1  0  0  5  0  0 32  1]\n",
      " [ 0  0  0  0  0  0  0  0  4  0  0  0  0  0  0 24]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix \n",
    "a = confusion_matrix(test_tag, pred) \n",
    "print a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rom pandas_confusion import ConfusionMatrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "delimiter = \"，|。|、|（|）|／|《|》|】|【|「|」|；|：|！\".decode('utf-8') \n",
    "def splitSentense(text, delimiter):\n",
    "    return re.split(delimiter, text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "content = allNews[3][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 364,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['From rpm-list-admin@freshrpms.net  Sat Oct  5 10:36:14 2002\\n', 'Return-Path: <rpm-zzzlist-admin@freshrpms.net>\\n', 'Delivered-To: yyyy@localhost.spamassassin.taint.org\\n', 'Received: from localhost (jalapeno [127.0.0.1])\\n', '\\tby jmason.org (Postfix) with ESMTP id 6C4D016F21\\n', '\\tfor <jm@localhost>; Sat,  5 Oct 2002 10:35:36 +0100 (IST)\\n', 'Received: from jalapeno [127.0.0.1]\\n', '\\tby localhost with IMAP (fetchmail-5.9.0)\\n', '\\tfor jm@localhost (single-drop); Sat, 05 Oct 2002 10:35:36 +0100 (IST)\\n', 'Received: from egwn.net (auth02.nl.egwn.net [193.172.5.4]) by\\n', '    dogma.slashnull.org (8.11.6/8.11.6) with ESMTP id g94NEAK13745 for\\n', '    <jm-rpm@jmason.org>; Sat, 5 Oct 2002 00:14:15 +0100\\n', 'Received: from auth02.nl.egwn.net (localhost [127.0.0.1]) by egwn.net\\n', '    (8.11.6/8.11.6/EGWN) with ESMTP id g94N82f30483; Sat, 5 Oct 2002 01:08:02\\n', '    +0200\\n', 'Received: from relay2.EECS.Berkeley.EDU (relay2.EECS.Berkeley.EDU\\n', '    [169.229.60.28]) by egwn.net (8.11.6/8.11.6/EGWN) with ESMTP id\\n', '    g94N7If23138 for <rpm-list@freshrpms.net>; Sat, 5 Oct 2002 01:07:18 +0200\\n', 'Received: from relay3.EECS.Berkeley.EDU (localhost.Berkeley.EDU\\n', '    [127.0.0.1]) by relay2.EECS.Berkeley.EDU (8.9.3/8.9.3) with ESMTP id\\n', '    QAA03058 for <rpm-list@freshrpms.net>; Fri, 4 Oct 2002 16:07:16 -0700\\n', '    (PDT)\\n', 'Received: from eecs.berkeley.edu (brawnix.CS.Berkeley.EDU [128.32.35.162])\\n', '    by relay3.EECS.Berkeley.EDU (8.9.3/8.9.3) with ESMTP id QAA16606 for\\n', '    <rpm-list@freshrpms.net>; Fri, 4 Oct 2002 16:07:11 -0700 (PDT)\\n', 'Message-Id: <3D9E1F20.3050300@eecs.berkeley.edu>\\n', 'From: Ben Liblit <liblit@eecs.berkeley.edu>\\n', 'User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1) Gecko/20020827\\n', 'X-Accept-Language: en-us, en\\n', 'MIME-Version: 1.0\\n', 'To: rpm-zzzlist@freshrpms.net\\n', 'Subject: Re: RedHat 8.0 and his own freetype\\n', 'References: <20021004155451.52f9ecd5.matthias_haase@bennewitz.com>\\n', 'Content-Type: text/plain; charset=us-ascii; format=flowed\\n', 'Content-Transfer-Encoding: 7bit\\n', 'X-Mailscanner: Found to be clean, Found to be clean\\n', 'Sender: rpm-zzzlist-admin@freshrpms.net\\n', 'Errors-To: rpm-zzzlist-admin@freshrpms.net\\n', 'X-Beenthere: rpm-zzzlist@freshrpms.net\\n', 'X-Mailman-Version: 2.0.11\\n', 'Precedence: bulk\\n', 'Reply-To: rpm-zzzlist@freshrpms.net\\n', 'List-Help: <mailto:rpm-zzzlist-request@freshrpms.net?subject=help>\\n', 'List-Post: <mailto:rpm-zzzlist@freshrpms.net>\\n', 'List-Subscribe: <http://lists.freshrpms.net/mailman/listinfo/rpm-zzzlist>,\\n', '    <mailto:rpm-list-request@freshrpms.net?subject=subscribe>\\n', 'List-Id: Freshrpms RPM discussion list <rpm-zzzlist.freshrpms.net>\\n', 'List-Unsubscribe: <http://lists.freshrpms.net/mailman/listinfo/rpm-zzzlist>,\\n', '    <mailto:rpm-list-request@freshrpms.net?subject=unsubscribe>\\n', 'List-Archive: <http://lists.freshrpms.net/pipermail/rpm-zzzlist/>\\n', 'X-Original-Date: Fri, 04 Oct 2002 16:07:12 -0700\\n', 'Date: Fri, 04 Oct 2002 16:07:12 -0700\\n', '\\n', 'Matthias Haase wrote:\\n', '> RH ships the code with the bytecode hinter disabled which makes \\n', '> non-AA fonts really ugly.\\n', '> This reqiures only a small change for include/freetype/config/ftoption.h,\\n', '> it is very well documented.\\n', '\\n', 'Red Hat 8.0 ships with the bytecode hinter enabled; I think 7.3 may have \\n', 'as well.\\n', '\\n', 'The enabling change to \"ftoption.h\" is made by Red Hat\\'s SRPM before \\n', 'building.  Take a look at \"freetype-2.1.1-enable-ft2-bci.patch\" from the \\n', \"SRPM; it's pretty clear that this does exactly what needs to be done.\\n\", '\\n', 'So if your fonts look ugly, lack of bytecode hinting is *not* the cause.\\n', '\\n', '\\n', '_______________________________________________\\n', 'RPM-List mailing list <RPM-List@freshrpms.net>\\n', 'http://lists.freshrpms.net/mailman/listinfo/rpm-list\\n', '\\n', '\\n']\n"
     ]
    }
   ],
   "source": [
    "f = open('data/easy_ham/01251.793e5c04967cb90191e805dfa619c55a', 'r')\n",
    "msg = f.readlines()\n",
    "print msg\n",
    "f.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 366,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Matthias Haase wrote:\n",
      "> RH ships the code with the bytecode hinter disabled which makes \n",
      "> non-AA fonts really ugly.\n",
      "> This reqiures only a small change for include/freetype/config/ftoption.h,\n",
      "> it is very well documented.\n",
      "\n",
      "Red Hat 8.0 ships with the bytecode hinter enabled; I think 7.3 may have \n",
      "as well.\n",
      "\n",
      "The enabling change to \"ftoption.h\" is made by Red Hat's SRPM before \n",
      "building.  Take a look at \"freetype-2.1.1-enable-ft2-bci.patch\" from the \n",
      "SRPM; it's pretty clear that this does exactly what needs to be done.\n",
      "\n",
      "So if your fonts look ugly, lack of bytecode hinting is *not* the cause.\n",
      "\n",
      "\n",
      "_______________________________________________\n",
      "RPM-List mailing list <RPM-List@freshrpms.net>\n",
      "http://lists.freshrpms.net/mailman/listinfo/rpm-list\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "first_blank_index = msg.index('\\n')\n",
    "#print first_blank_index\n",
    "msg = msg[(first_blank_index + 1): ]\n",
    "print ''.join(msg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 367,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Matthias Haase wrote:\n",
      "> RH ships the code with the bytecode hinter disabled which makes \n",
      "> non-AA fonts really ugly.\n",
      "> This reqiures only a small change for include/freetype/config/ftoption.h,\n",
      "> it is very well documented.\n",
      "\n",
      "Red Hat 8.0 ships with the bytecode hinter enabled; I think 7.3 may have \n",
      "as well.\n",
      "\n",
      "The enabling change to \"ftoption.h\" is made by Red Hat's SRPM before \n",
      "building.  Take a look at \"freetype-2.1.1-enable-ft2-bci.patch\" from the \n",
      "SRPM; it's pretty clear that this does exactly what needs to be done.\n",
      "\n",
      "So if your fonts look ugly, lack of bytecode hinting is *not* the cause.\n",
      "\n",
      "\n",
      "_______________________________________________\n",
      "RPM-List mailing list <RPM-List@freshrpms.net>\n",
      "http://lists.freshrpms.net/mailman/listinfo/rpm-list\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "def get_msg(path):    \n",
    "    f = open(path, 'r')\n",
    "    msg = f.readlines()\n",
    "    f.close()\n",
    "    first_blank_index = msg.index('\\n')\n",
    "    msg = msg[(first_blank_index + 1): ]\n",
    "    return ''.join(msg)\n",
    "\n",
    "print get_msg('data/easy_ham/01251.793e5c04967cb90191e805dfa619c55a')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 369,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import os\n",
    "filelist = os.listdir('data/easy_ham')\n",
    "#print filelist\n",
    "\n",
    "filelist = [f for f in filelist if f != 'cmds']\n",
    "#print filelist\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_msgdir(path):\n",
    "    filelist = os.listdir(path)\n",
    "    filelist = [f for f in filelist if f != 'cmds']\n",
    "    all_msgs = [get_msg(os.path.join(path, f)) for f in filelist]\n",
    "    return all_msgs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 371,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#讀取訓練資料集\n",
    "train_spam_messages    = get_msgdir('data/spam')\n",
    "train_easyham_messages = get_msgdir('data/easy_ham')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 372,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#讀取測試資料集\n",
    "test_spam_messages    = get_msgdir('data/spam_2')\n",
    "test_easyham_messages = get_msgdir('data/easy_ham_2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 375,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Martin A posted:\n",
      "Tassos Papadopoulos, the Greek sculptor behind the plan, judged that the\n",
      " limestone of Mount Kerdylio, 70 miles east of Salonika and not far from the\n",
      " Mount Athos monastic community, was ideal for the patriotic sculpture. \n",
      " \n",
      " As well as Alexander's granite features, 240 ft high and 170 ft wide, a\n",
      " museum, a restored amphitheatre and car park for admiring crowds are\n",
      "planned\n",
      "---------------------\n",
      "So is this mountain limestone or granite?\n",
      "If it's limestone, it'll weather pretty fast.\n",
      "\n",
      "------------------------ Yahoo! Groups Sponsor ---------------------~-->\n",
      "4 DVDs Free +s&p Join Now\n",
      "http://us.click.yahoo.com/pt6YBB/NXiEAA/mG3HAA/7gSolB/TM\n",
      "---------------------------------------------------------------------~->\n",
      "\n",
      "To unsubscribe from this group, send an email to:\n",
      "forteana-unsubscribe@egroups.com\n",
      "\n",
      " \n",
      "\n",
      "Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ \n",
      "\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#train_spam_messages[1]\n",
    "print train_easyham_messages[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 407,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_msg_words(msg, stopwords = []):\n",
    "    msg = re.sub('3D', '', msg)\n",
    "    \n",
    "    msg = re.sub('<(.|\\n)*?>', ' ', msg)\n",
    "    msg = re.sub('&\\w+;', ' ', msg)\n",
    "\n",
    "    msg = re.sub('_+', '_', msg)\n",
    "    msg_words = set(wordpunct_tokenize(msg.replace('=\\n', '').lower()))\n",
    "\n",
    "    msg_words = msg_words.difference(stopwords)\n",
    "    msg_words = [w for w in msg_words if re.search('[a-zA-Z]', w) and len(w) > 1]\n",
    "    return msg_words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 408,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['code', 'matthias', 'lack', 'disabled', 'done', 'bytecode', 'config', 'rpm', 'your', 'before', 'mailman', 'ships', 'to', 'only', 'pretty', 'include', 'hat', 'building', 'freetype', 'enable', 'which', 'that', 'very', 'exactly', 'lists', 'made', 'not', 'documented', 'freshrpms', 'with', 'bci', 'look', 'this', 'fonts', 'list', 'patch', 'ugly', 'reqiures', 'small', 'srpm', 'clear', 'wrote', 'aa', 'haase', 'enabled', 'is', 'it', 'hinter', 'as', 'at', 'have', 'take', 'really', 'what', 'mailing', 'does', 'listinfo', 'ftoption', 'net', 'if', 'rh', 'red', 'be', 'http', 'may', 'needs', 'from', 'by', 'change', 'ft2', 'for', 'of', 'hinting', 'well', 'non', 'enabling', 'so', 'the', 'makes', 'think', 'cause']\n"
     ]
    }
   ],
   "source": [
    "import re\n",
    "from nltk.tokenize import wordpunct_tokenize\n",
    "msg = get_msg('data/easy_ham/01251.793e5c04967cb90191e805dfa619c55a')\n",
    "print get_msg_words(msg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 409,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'i', u'me', u'my', u'myself', u'we', u'our', u'ours', u'ourselves', u'you', u'your', u'yours', u'yourself', u'yourselves', u'he', u'him', u'his', u'himself', u'she', u'her', u'hers', u'herself', u'it', u'its', u'itself', u'they', u'them', u'their', u'theirs', u'themselves', u'what', u'which', u'who', u'whom', u'this', u'that', u'these', u'those', u'am', u'is', u'are', u'was', u'were', u'be', u'been', u'being', u'have', u'has', u'had', u'having', u'do', u'does', u'did', u'doing', u'a', u'an', u'the', u'and', u'but', u'if', u'or', u'because', u'as', u'until', u'while', u'of', u'at', u'by', u'for', u'with', u'about', u'against', u'between', u'into', u'through', u'during', u'before', u'after', u'above', u'below', u'to', u'from', u'up', u'down', u'in', u'out', u'on', u'off', u'over', u'under', u'again', u'further', u'then', u'once', u'here', u'there', u'when', u'where', u'why', u'how', u'all', u'any', u'both', u'each', u'few', u'more', u'most', u'other', u'some', u'such', u'no', u'nor', u'not', u'only', u'own', u'same', u'so', u'than', u'too', u'very', u's', u't', u'can', u'will', u'just', u'don', u'should', u'now', 'll', 've']\n"
     ]
    }
   ],
   "source": [
    "from nltk.corpus import stopwords\n",
    "sw = stopwords.words('english')\n",
    "sw.extend(['ll', 've'])\n",
    "print sw"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 410,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'aa': True, 'the': True, 'code': True, 'matthias': True, 'enabled': True, 'is': True, 'lack': True, 'it': True, 'disabled': True, 'really': True, 'as': True, 'done': True, 'bytecode': True, 'have': True, 'rpm': True, 'your': True, 'before': True, 'what': True, 'rh': True, 'mailman': True, 'for': True, 'may': True, 'haase': True, 'ships': True, 'to': True, 'only': True, 'does': True, 'listinfo': True, 'take': True, 'which': True, 'so': True, 'net': True, 'include': True, 'hat': True, 'red': True, 'ftoption': True, 'be': True, 'freetype': True, 'enable': True, 'http': True, 'pretty': True, 'that': True, 'very': True, 'exactly': True, 'needs': True, 'lists': True, 'well': True, 'non': True, 'ft2': True, 'if': True, 'makes': True, 'not': True, 'from': True, 'documented': True, 'freshrpms': True, 'with': True, 'by': True, 'change': True, 'look': True, 'building': True, 'hinter': True, 'think': True, 'made': True, 'bci': True, 'this': True, 'config': True, 'clear': True, 'fonts': True, 'list': True, 'patch': True, 'enabling': True, 'ugly': True, 'cause': True, 'reqiures': True, 'of': True, 'small': True, 'srpm': True, 'mailing': True, 'hinting': True, 'wrote': True, 'at': True}\n"
     ]
    }
   ],
   "source": [
    "features = {}\n",
    "msg_words = get_msg_words(msg)\n",
    "#print msg_words\n",
    "\n",
    "for  w in msg_words:\n",
    "    features[w] = True\n",
    "\n",
    "print features\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 411,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def word_indicator(msg):\n",
    "    features = {}\n",
    "    msg_words = get_msg_words(msg)\n",
    "    for  w in msg_words:\n",
    "        features[w] = True\n",
    "    return features\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 412,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def features_from_messages(messages, label, feature_extractor):\n",
    "    features_labels = []\n",
    "    for msg in messages:\n",
    "        features = feature_extractor(msg)\n",
    "        features_labels.append((features, label))\n",
    "    return features_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 413,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_spam = features_from_messages(train_spam_messages, 'spam', \n",
    "                                        word_indicator)\n",
    "train_ham = features_from_messages(train_easyham_messages, 'ham', \n",
    "                                       word_indicator)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 421,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#print train_spam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 415,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_set = train_spam + train_ham"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 416,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test_spam = features_from_messages(test_spam_messages, 'spam',\n",
    "                                       word_indicator)\n",
    "\n",
    "test_ham = features_from_messages(test_easyham_messages, 'ham',\n",
    "                                     word_indicator)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 417,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from nltk import NaiveBayesClassifier\n",
    "classifier = NaiveBayesClassifier.train(train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 418,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Spam accuracy: 97.21%\n",
      "Test Ham accuracy: 87.92%\n"
     ]
    }
   ],
   "source": [
    "print ('Test Spam accuracy: {0:.2f}%'\n",
    "  .format(100 * nltk.classify.accuracy(classifier, test_spam)))\n",
    "\n",
    "print ('Test Ham accuracy: {0:.2f}%'\n",
    "  .format(100 * nltk.classify.accuracy(classifier, test_ham)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 419,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Most Informative Features\n",
      "                mailings = True             spam : ham    =    124.8 : 1.0\n",
      "                unwanted = True             spam : ham    =     71.6 : 1.0\n",
      "              requesting = True             spam : ham    =     64.9 : 1.0\n",
      "                mortgage = True             spam : ham    =     64.9 : 1.0\n",
      "            unsubscribed = True             spam : ham    =     62.9 : 1.0\n",
      "                 removal = True             spam : ham    =     56.0 : 1.0\n",
      "               sincerely = True             spam : ham    =     49.2 : 1.0\n",
      "             instruction = True             spam : ham    =     48.3 : 1.0\n",
      "         confidentiality = True             spam : ham    =     48.3 : 1.0\n",
      "                  kindly = True             spam : ham    =     46.9 : 1.0\n",
      "                   opted = True             spam : ham    =     44.9 : 1.0\n",
      "                      pa = True             spam : ham    =     44.9 : 1.0\n",
      "                     sir = True             spam : ham    =     44.9 : 1.0\n",
      "              profitable = True             spam : ham    =     42.1 : 1.0\n",
      "              instructed = True             spam : ham    =     41.6 : 1.0\n",
      "              subscriber = True             spam : ham    =     41.6 : 1.0\n",
      "                      mo = True             spam : ham    =     41.6 : 1.0\n",
      "             subscribers = True             spam : ham    =     41.6 : 1.0\n",
      "                  postal = True             spam : ham    =     41.6 : 1.0\n",
      "              assistance = True             spam : ham    =     40.4 : 1.0\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "print classifier.show_most_informative_features(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 459,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import sqlite3\n",
    "db = sqlite3.connect('news2.sqlite') \n",
    "cur = db.cursor()\n",
    "cur.execute('select title, summary, category from news_entry')\n",
    "allNews = cur.fetchall()\n",
    "\n",
    "corpus = []\n",
    "tags = []\n",
    "title = []\n",
    "for rec in allNews:\n",
    "    #if (rec[2] == '娛樂'.decode('utf-8')):\n",
    "    ele  = jieba.cut(rec[1].strip())\n",
    "    corpus.append([w for w in ' '.join(ele).split()])\n",
    "cur.close()\n",
    "db.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 460,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from gensim.models import word2vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 461,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:gensim.models.word2vec:consider setting layer size to a multiple of 4 for greater performance\n"
     ]
    }
   ],
   "source": [
    "num_features = 10\n",
    "min_word_count = 10\n",
    "num_workers = 4\n",
    "context = 10\n",
    "epoch = 20\n",
    "sample = 1e-5\n",
    "model = word2vec.Word2Vec(corpus, workers = num_workers,\n",
    "                          sample = sample,\n",
    "                          size = num_features,\n",
    "                          min_count=min_word_count,\n",
    "                          window = context,\n",
    "                          iter = epoch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 462,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "西方\n",
      "第二\n",
      "下周\n",
      "新生\n",
      "反對\n",
      "創意\n",
      "舉辦\n",
      "怎麼\n",
      "作弊\n",
      "造\n",
      "錢\n",
      "影音\n",
      "之間\n",
      "案件\n",
      "下滑\n",
      "搜尋\n",
      "欲\n",
      "成分\n",
      "此外\n",
      "行駛\n",
      "傍晚\n",
      "@\n",
      "食物\n",
      "談\n",
      "許多\n",
      "反\n",
      "品牌\n",
      "記者\n",
      "那裡\n",
      "過程\n",
      "質詢\n",
      "臉書上\n",
      "遠\n",
      "除\n",
      "罪\n",
      "貓\n",
      "歲\n",
      "猿隊\n",
      "其\n",
      "接近\n",
      "對此\n",
      "態度\n",
      "第一\n",
      "紛紛\n",
      "商務\n",
      "個\n",
      "第三\n",
      "分布\n",
      "交往\n",
      "朱立\n",
      "委員\n",
      "無\n",
      "歐巴馬\n",
      "消防\n",
      "公路\n",
      "+\n",
      "三年\n",
      "是\n",
      "Z5\n",
      "台灣黃蝶\n",
      "香水\n",
      "歐洲\n",
      "多名\n",
      "人潮\n",
      "兄弟\n",
      "兩天\n",
      "明顯\n",
      "五年\n",
      "在家\n",
      "戰機\n",
      "大批\n",
      "嗎\n",
      "適用\n",
      "發展\n",
      "找到\n",
      "教\n",
      "100\n",
      "也\n",
      "入圍\n",
      "看來\n",
      "以來\n",
      "警局\n",
      "Line\n",
      "最近\n",
      "併\n",
      "執導\n",
      "人數\n",
      "基\n",
      "三星\n",
      "官員\n",
      "目前\n",
      "月初\n",
      "上\n",
      "劍\n",
      "拿到\n",
      "他們\n",
      "經\n",
      "屋主\n",
      "考量\n",
      "女方\n",
      "禁止\n",
      "受到\n",
      "李女\n",
      "男友\n",
      "丈夫\n",
      "表明\n",
      "穿越\n",
      "台中市\n",
      "大眾\n",
      "覺得\n",
      "小S\n",
      "這場\n",
      "依法\n",
      "俄國\n",
      "白色\n",
      "國中\n",
      "‧\n",
      "議會\n",
      "分心\n",
      "電腦\n",
      "公民\n",
      "90\n",
      "醫生\n",
      "正式\n",
      "事項\n",
      "以下\n",
      "以上\n",
      "相對\n",
      "柯\n",
      "利益\n",
      "主要\n",
      "該\n",
      "做出\n",
      "水管\n",
      "任務\n",
      "老\n",
      "明天\n",
      "合宜\n",
      "圖片\n",
      "這句\n",
      "in\n",
      "智慧\n",
      "這\n",
      "起訴\n",
      "本土\n",
      "想要\n",
      "階段\n",
      "陸續\n",
      "冠\n",
      "中信\n",
      "清晨\n",
      "鬧\n",
      "Journal\n",
      "條款\n",
      "配合\n",
      "第一次\n",
      "課\n",
      "內政部\n",
      "一段\n",
      "吸引\n",
      "A\n",
      "派\n",
      "旅車\n",
      "全台\n",
      "after\n",
      "前\n",
      "勞動\n",
      "當地\n",
      "駐\n",
      "特偵組\n",
      "轟\n",
      "意見\n",
      "加上\n",
      "不同\n",
      "......\n",
      "不斷\n",
      "適合\n",
      "正常\n",
      "不吐\n",
      "李詩欽\n",
      "連結\n",
      "責任\n",
      "台中\n",
      "不知\n",
      "醫療\n",
      "再\n",
      "「\n",
      "派出所\n",
      "看起\n",
      "汙染\n",
      "祝福\n",
      "癌症\n",
      "一步\n",
      "經濟\n",
      "連續\n",
      "宣布\n",
      "此事\n",
      "爆\n",
      ",\n",
      "模式\n",
      "印象\n",
      "水\n",
      "400\n",
      "町\n",
      "所\n",
      "犧牲\n",
      "首次\n",
      "薪水\n",
      "都市\n",
      "重\n",
      "が\n",
      "使用者\n",
      "9\n",
      "黨部\n",
      "足球\n",
      "基地\n",
      "透露\n",
      "牠\n",
      "難\n",
      "緊張\n",
      "違法\n",
      "APP\n",
      "貿易\n",
      "東西\n",
      "聯賽\n",
      "發布\n",
      "突然\n",
      "關注\n",
      "像是\n",
      "故意\n",
      "錯誤\n",
      "女星\n",
      "讀者\n",
      "載\n",
      "越\n",
      "れ\n",
      "每次\n",
      "笑\n",
      "爸爸\n",
      "前進\n",
      "攝影師\n",
      "提早\n",
      "居民\n",
      "一條\n",
      "goo\n",
      "法官\n",
      "到場\n",
      "自造\n",
      "富商\n",
      "打開\n",
      "誰\n",
      "薪資\n",
      "部分\n",
      "躺\n",
      "中國大陸\n",
      "懂\n",
      "拿出\n",
      "家屬\n",
      "Elsevier\n",
      "決策\n",
      "分局\n",
      "全部\n",
      "秒\n",
      "獅子\n",
      "王金平\n",
      "不料\n",
      "獨家\n",
      "獨立\n",
      "兩個\n",
      "市政府\n",
      "新知\n",
      "相\n",
      "好\n",
      "什麼\n",
      "青年\n",
      "王\n",
      "吃\n",
      "放在\n",
      "里長\n",
      "或\n",
      "條\n",
      "快來\n",
      "晚上\n",
      "這種\n",
      "舞台\n",
      "本季\n",
      "預期\n",
      "行動\n",
      "遭到\n",
      "相關\n",
      "意識\n",
      "》\n",
      "B\n",
      "資金\n",
      "佈\n",
      "兌\n",
      "子\n",
      "英格蘭\n",
      "超標\n",
      "毛\n",
      "加油\n",
      "肇事\n",
      "跟\n",
      "查看\n",
      "亞洲\n",
      "應該\n",
      "雕塑\n",
      "心理\n",
      "赴\n",
      "帶著\n",
      "證實\n",
      "會\n",
      "擔任\n",
      "複\n",
      "三\n",
      "總統候\n",
      "」\n",
      "人氣\n",
      "營運\n",
      "好萊塢\n",
      "相當\n",
      "特殊\n",
      "無法\n",
      "大\n",
      "-\n",
      "市場\n",
      "還好\n",
      "形成\n",
      "趴\n",
      "看法\n",
      "學習\n",
      "察院\n",
      "世紀\n",
      "影響\n",
      "低氣壓\n",
      "里\n",
      "問\n",
      "積極\n",
      "網址\n",
      "拖\n",
      "據\n",
      "新竹市\n",
      "竟然\n",
      "基隆\n",
      "家族\n",
      "行程\n",
      "話\n",
      "QE\n",
      "追問\n",
      "協會\n",
      "一一\n",
      "紀錄\n",
      "到底\n",
      "一下\n",
      "權\n",
      "成功\n",
      "收到\n",
      "迅速\n",
      "年前\n",
      "看護\n",
      "思考\n",
      "委員會\n",
      "現象\n",
      "真的\n",
      "依\n",
      "訂定\n",
      "新北\n",
      "起來\n",
      "那\n",
      "台\n",
      "徒刑\n",
      "座\n",
      "鑑定\n",
      "上網\n",
      "工時\n",
      "漲\n",
      "投手\n",
      "俄羅斯\n",
      "拍攝\n",
      "片中\n",
      "裝置\n",
      "中國國家\n",
      "旁\n",
      "預計\n",
      "案\n",
      "秘密\n",
      "煎\n",
      "排除\n",
      "客戶\n",
      "歐蘭德\n",
      "也將\n",
      "同時\n",
      "健康\n",
      "新光\n",
      "逮捕\n",
      "條件\n",
      "真相\n",
      "不再\n",
      "結果\n",
      "睡眠\n",
      "行經\n",
      "當\n",
      "選\n",
      "環景\n",
      "運作\n",
      "子女\n",
      "狀況\n",
      "一事\n",
      "變\n",
      "撞擊\n",
      "住宅\n",
      "一些\n",
      "Android\n",
      "John\n",
      "支安打\n",
      "宣稱\n",
      "事發\n",
      "02\n",
      "連\n",
      "said\n",
      "07\n",
      "人權\n",
      "天\n",
      "風\n",
      "08\n",
      "09\n",
      "企業\n",
      "結束\n",
      "直播\n",
      "答案\n",
      "一人\n",
      "自來\n",
      "女性\n",
      "運\n",
      "360\n",
      "萬元\n",
      "車輛\n",
      "股票\n",
      "前天\n",
      "專業\n",
      "其中\n",
      "想法\n",
      "不用\n",
      "不然\n",
      "卡麥隆\n",
      "即時\n",
      "機\n",
      "左\n",
      "手機\n",
      "罷免\n",
      "一件\n",
      "嗶\n",
      "方向\n",
      "com\n",
      "當初\n",
      "穿\n",
      "一份\n",
      "改革\n",
      "─\n",
      "北京\n",
      "這對\n",
      "訴訟\n",
      "擁有\n",
      "開\n",
      "訊\n",
      "還要\n",
      "令人\n",
      "英文\n",
      "該名\n",
      "市府\n",
      "一輛\n",
      "條例\n",
      "不僅\n",
      "透過\n",
      "能源\n",
      "館\n",
      "建商\n",
      "元素\n",
      "本\n",
      ".\n",
      "全國\n",
      "顧問\n",
      "陳\n",
      "16V\n",
      "刺\n",
      "突發\n",
      "時\n",
      "反正\n",
      "研究\n",
      "砍\n",
      "台語\n",
      "疑似\n",
      "危機\n",
      "過\n",
      "姓\n",
      "全面\n",
      "康熙\n",
      "作\n",
      "盤\n",
      "員\n",
      "小王子\n",
      "問題\n",
      "地檢署\n",
      "工會\n",
      "初步\n",
      "走\n",
      "規定\n",
      "這項\n",
      "酸\n",
      "TDW\n",
      "和平\n",
      "代表\n",
      "期望\n",
      "is\n",
      "it\n",
      "多元\n",
      "少\n",
      "保密\n",
      "取得\n",
      "’\n",
      "這位\n",
      "id\n",
      "新闢\n",
      "黨團\n",
      "平常\n",
      "別\n",
      "下車\n",
      "不過\n",
      "海上\n",
      "土地\n",
      "契約\n",
      "主人\n",
      "罵\n",
      "總\n",
      "東京\n",
      "身份\n",
      "捲\n",
      "目的\n",
      "部長\n",
      "至今\n",
      "大聯盟\n",
      "VS\n",
      "指標\n",
      "車上\n",
      "SONY\n",
      "on\n",
      "一系列\n",
      "畫面\n",
      "離\n",
      "策略\n",
      "首度\n",
      "達成\n",
      "訊息\n",
      "經紀\n",
      "臨時\n",
      "安倍\n",
      "the\n",
      "有關\n",
      "儀\n",
      "主任\n",
      "不可\n",
      "包包\n",
      "變成\n",
      "同\n",
      "大多\n",
      "裡面\n",
      "一路\n",
      "檢署\n",
      "熱烈\n",
      "天氣\n",
      "離開\n",
      "附近\n",
      "產\n",
      "還說\n",
      "太\n",
      "had\n",
      "萬\n",
      "不及\n",
      "你們\n",
      "文章\n",
      "全球\n",
      "就業\n",
      "尼\n",
      "但\n",
      "發稿\n",
      "感動\n",
      "楊\n",
      "iOS\n",
      "像\n",
      "駕\n",
      "美好\n",
      "過失\n",
      "55\n",
      "54\n",
      "56\n",
      "51\n",
      "50\n",
      "遊戲\n",
      "籃板\n",
      "隨後\n",
      "質疑\n",
      "刑事\n",
      "引發\n",
      "www\n",
      "現任\n",
      "來\n",
      "月間\n",
      "陷入\n",
      "』\n",
      "學童\n",
      "商品\n",
      "使用\n",
      "折\n",
      "答詢\n",
      "更好\n",
      "for\n",
      "性感\n",
      "痛批\n",
      "傷害\n",
      "這群\n",
      "/\n",
      "感到\n",
      "數據\n",
      "律師\n",
      "法律\n",
      "打出\n",
      "更多\n",
      "程度\n",
      "在網\n",
      "畢業\n",
      "不錯\n",
      "完全\n",
      "車\n",
      "就要\n",
      "回應\n",
      "補充\n",
      "著\n",
      "明年\n",
      "地位\n",
      "睡\n",
      "因\n",
      "作業\n",
      "入\n",
      "發言人\n",
      "並無\n",
      "看見\n",
      "主委\n",
      "程序\n",
      "聯\n",
      "否認\n",
      "有用\n",
      "鉛管\n",
      "兩名\n",
      "施旖婕\n",
      "成員\n",
      "並將\n",
      "勸\n",
      "出\n",
      "最\n",
      "總統\n",
      "同樣\n",
      "為度\n",
      "各項\n",
      "累積\n",
      "大方\n",
      "供\n",
      "打造\n",
      "合約\n",
      "據悉\n",
      "解釋\n",
      "颱\n",
      "警官\n",
      "下午\n",
      "小時\n",
      "安全\n",
      "理財\n",
      "was\n",
      "中小\n",
      "近年\n",
      "請求\n",
      "感受\n",
      "性能\n",
      "量\n",
      "和\n",
      "黑\n",
      "主辦\n",
      "旗\n",
      "周董\n",
      "睡覺\n",
      "顏女\n",
      "詞\n",
      "盡\n",
      "靠\n",
      "口\n",
      "呢\n",
      "駕駛\n",
      "送醫\n",
      "是不是\n",
      "鬼\n",
      "受傷\n",
      "一起\n",
      "翻\n",
      "究竟\n",
      "文字\n",
      "哪\n",
      "對象\n",
      "山林\n",
      "了\n",
      "交給\n",
      "看\n",
      "方面\n",
      "名\n",
      "下班\n",
      "涉及\n",
      "談會\n",
      "習慣\n",
      "社群\n",
      "上市\n",
      "受\n",
      "度\n",
      "搭\n",
      "鐵路\n",
      "這麼\n",
      "男\n",
      "顯然\n",
      "為\n",
      "隔\n",
      "左右\n",
      "勞動部\n",
      "克斯\n",
      "溝通\n",
      "贏得\n",
      "有名\n",
      "肯定\n",
      "敗\n",
      "下去\n",
      "學家\n",
      "卻\n",
      "法廣\n",
      "朱立倫\n",
      "追查\n",
      "撞\n",
      "產生\n",
      "查出\n",
      "申報\n",
      "改善\n",
      "被害人\n",
      "雙方\n",
      "技職\n",
      "長\n",
      "棒球\n",
      "設施\n",
      "曾\n",
      "通報\n",
      "佔\n",
      "要\n",
      "票券\n",
      "反映\n",
      "水果\n",
      "外野\n",
      "塔柱\n",
      "【\n",
      "台北市\n",
      "趙\n",
      "分隊\n",
      "優惠\n",
      "必須\n",
      "日前\n",
      "美國\n",
      "課綱\n",
      "藉由\n",
      "通知\n",
      "選區\n",
      "表演\n",
      "專訪\n",
      "0\n",
      "罕見\n",
      "報警\n",
      "隨著\n",
      "業績\n",
      "服務\n",
      "沒有\n",
      "各地\n",
      "最終\n",
      "科罰\n",
      "再次\n",
      "分鐘\n",
      "喜歡\n",
      "面臨\n",
      "設計\n",
      "心情\n",
      "儀式\n",
      "金\n",
      "警員\n",
      "就會\n",
      "是否\n",
      "八仙\n",
      "加班\n",
      "國會\n",
      "審團\n",
      "周\n",
      "平台\n",
      "創業\n",
      "意思\n",
      "園\n",
      "主持人\n",
      "民事\n",
      "老人\n",
      "二壘\n",
      "市議員\n",
      "少年\n",
      "與\n",
      "一把\n",
      "早已\n",
      "因為\n",
      "民進\n",
      "當年\n",
      "價格\n",
      "樂園\n",
      "清潔\n",
      "創造\n",
      "回來\n",
      "沈啟\n",
      "客\n",
      "大樓\n",
      "試圖\n",
      "塑化\n",
      "跳\n",
      "等\n",
      "名字\n",
      "訪問\n",
      "所幸\n",
      "偵辦\n",
      "金融\n",
      "午餐\n",
      "阿\n",
      "貼圖\n",
      "保障\n",
      "民主\n",
      "年底\n",
      "老虎\n",
      "嫌\n",
      "意外\n",
      "蘇女\n",
      "[\n",
      "多少\n",
      "號\n",
      "我們\n",
      "公斤\n",
      "火警\n",
      "正確\n",
      "穿著\n",
      "提供\n",
      "出來\n",
      "爆發\n",
      "父母\n",
      "現身\n",
      "更\n",
      "(\n",
      "年度\n",
      "成立\n",
      "警告\n",
      "保險\n",
      "推出\n",
      "送\n",
      "這個\n",
      "事實\n",
      "狀態\n",
      "每股\n",
      "間\n",
      "性交\n",
      "育成\n",
      "搞\n",
      "換柱\n",
      "習近\n",
      "用\n",
      "新聞時\n",
      "自我\n",
      "指數\n",
      "移送\n",
      "周日\n",
      "司法\n",
      "需求\n",
      "果然\n",
      "研發\n",
      "正在\n",
      "全世界\n",
      "F\n",
      "早就\n",
      "銷售\n",
      "民航局\n",
      "度假\n",
      "住院\n",
      "依據\n",
      "布瑞南\n",
      "心臟\n",
      "週刊\n",
      "行為\n",
      "緣\n",
      "て\n",
      "發生\n",
      "陪\n",
      "手術\n",
      "學生\n",
      "官網\n",
      "未來\n",
      "高度\n",
      "瞬間\n",
      "清\n",
      "指\n",
      "完\n",
      "／\n",
      "】\n",
      "爭取\n",
      "當中\n",
      "は\n",
      "安心\n",
      "休息\n",
      "購物\n",
      "文長\n",
      "廣告\n",
      "電視\n",
      "南部\n",
      "1\n",
      "臉\n",
      "已\n",
      "男子\n",
      "每日\n",
      "!\n",
      "金額\n",
      "久\n",
      "釐\n",
      "太太\n",
      "菜\n",
      "接下\n",
      "精神\n",
      "城市\n",
      "吳男\n",
      "本人\n",
      "做好\n",
      "不少\n",
      "難民\n",
      "NCC\n",
      "名單\n",
      "超\n",
      "學院\n",
      "全場\n",
      "北投\n",
      "兩岸\n",
      "爸媽\n",
      "桃園\n",
      "理事\n",
      "“\n",
      "預測\n",
      "管\n",
      "女人\n",
      "夢\n",
      "police\n",
      "產品\n",
      "教學\n",
      "推動\n",
      "戰\n",
      "確定\n",
      "常\n",
      "越南\n",
      "變化\n",
      "作用\n",
      "旅行\n",
      "各界\n",
      "確實\n",
      "蕊\n",
      "家庭\n",
      "專案\n",
      "兒\n",
      "更多國際\n",
      "歷史\n",
      "not\n",
      "橘子\n",
      "特色\n",
      "普拉斯\n",
      "醫師\n",
      "評論\n",
      "上班\n",
      "造成\n",
      "恐嚇\n",
      "勞基法\n",
      "と\n",
      "回收\n",
      "街頭\n",
      "負面\n",
      "副刊\n",
      "有\n",
      "隨即\n",
      "父親\n",
      "採訪\n",
      "小\n",
      "美\n",
      "Live\n",
      "稿酬\n",
      "機車\n",
      "件\n",
      "別人\n",
      "學者\n",
      "家人\n",
      "家長\n",
      "瓦斯\n",
      "當局\n",
      "還是\n",
      "縣\n",
      "建立\n",
      "生父\n",
      "詳細\n",
      "所屬\n",
      "曝光\n",
      "毛治國\n",
      "家\n",
      "多年\n",
      "投資\n",
      "今\n",
      "咖啡\n",
      "局上\n",
      "局下\n",
      "堅持\n",
      "唯一\n",
      "上午\n",
      "街\n",
      "武器\n",
      "打工\n",
      "全壘\n",
      "で\n",
      "另\n",
      "App\n",
      "業\n",
      "修\n",
      "交通\n",
      "失誤\n",
      "生物\n",
      "寶\n",
      "重點\n",
      "1000\n",
      "工業\n",
      "預估\n",
      "月\n",
      "男學生\n",
      "保留\n",
      "減少\n",
      "難以\n",
      "劑\n",
      "倒\n",
      "直擊\n",
      "報案\n",
      "疑慮\n",
      "那個\n",
      "之一\n",
      "演講\n",
      "之下\n",
      "經過\n",
      "患者\n",
      "花\n",
      "鄉\n",
      "2\n",
      "大概\n",
      "去\n",
      "花蓮\n",
      "that\n",
      "外界\n",
      "列車\n",
      "二\n",
      "11\n",
      "10\n",
      "13\n",
      "12\n",
      "鳥\n",
      "14\n",
      "17\n",
      "16\n",
      "19\n",
      "18\n",
      "路\n",
      "聲\n",
      "立法\n",
      "were\n",
      "社區\n",
      "都\n",
      "兼\n",
      "and\n",
      "讀\n",
      "較\n",
      "新店\n",
      "希特勒\n",
      "還有\n",
      "澄清\n",
      "世界\n",
      "實施\n",
      "宣傳\n",
      "去年同期\n",
      "を\n",
      "男女\n",
      "喔\n",
      "先前\n",
      "飯店\n",
      "估計\n",
      "”\n",
      "25\n",
      "成長\n",
      "讚\n",
      "農田\n",
      "建築\n",
      "ㄧ\n",
      "參選人\n",
      "登記\n",
      "韓國\n",
      "幾乎\n",
      "迴\n",
      "200\n",
      "中文\n",
      "監測\n",
      "集團\n",
      "用戶\n",
      "適當\n",
      "舉行\n",
      "上前\n",
      "作品\n",
      "發出\n",
      "]\n",
      "移動\n",
      "管理\n",
      "除了\n",
      "好友\n",
      "幹\n",
      "接著\n",
      "保單\n",
      "中方\n",
      "原因\n",
      "趁\n",
      "shot\n",
      "軟體\n",
      "成績\n",
      "大人\n",
      "彼此\n",
      "身材\n",
      "新手\n",
      "食品\n",
      "各\n",
      "地產\n",
      "患有\n",
      "兒子\n",
      "最大\n",
      "使得\n",
      "同意\n",
      "地方\n",
      "辭\n",
      "巴勒斯坦\n",
      "塞德\n",
      "戴\n",
      "則是\n",
      "砸\n",
      "網友\n",
      "不妥\n",
      "只好\n",
      "政治\n",
      "掌握\n",
      "應\n",
      "投訴\n",
      "享有\n",
      "民調\n",
      "一處\n",
      "例如\n",
      "基礎\n",
      "啦\n",
      "早餐\n",
      "打字\n",
      "只要\n",
      "想\n",
      "還能\n",
      "直接\n",
      "發\n",
      "勿\n",
      "最好\n",
      "助攻\n",
      "需要\n",
      "帳號\n",
      "路邊\n",
      "－\n",
      "兩年\n",
      "係\n",
      "表達\n",
      "不好\n",
      "警察\n",
      "LINE\n",
      "竄\n",
      "吳思華\n",
      "阿里山\n",
      "高中\n",
      "朋友\n",
      "總經理\n",
      "球員\n",
      "一家\n",
      "打算\n",
      "因而\n",
      "3\n",
      "分別\n",
      "玩樂\n",
      "利用\n",
      "藍鳥\n",
      "位置\n",
      "限制\n",
      "先生\n",
      "傷亡\n",
      "醫院\n",
      "js\n",
      "苗栗\n",
      "前年\n",
      "道\n",
      "過去\n",
      "道路\n",
      "批\n",
      "工人\n",
      "平均\n",
      "情緒\n",
      "appledaily\n",
      "生命\n",
      "一次\n",
      "邱男\n",
      "幾天\n",
      "兩\n",
      "部\n",
      "遇到\n",
      "こ\n",
      "╱\n",
      "玉山\n",
      "s\n",
      "參選\n",
      "標準\n",
      "她\n",
      "而已\n",
      "容易\n",
      "問金\n",
      "億英\n",
      "請連\n",
      "重大\n",
      "這名\n",
      "故事\n",
      "有點\n",
      "肉\n",
      "愈\n",
      "舊\n",
      "，\n",
      "60\n",
      "下降\n",
      "68\n",
      "者\n",
      "研判\n",
      "分手\n",
      "前往\n",
      "數位\n",
      "刪\n",
      "亭\n",
      "盜\n",
      "工作\n",
      "周三\n",
      "專輯\n",
      "德\n",
      "答應\n",
      "國\n",
      "500\n",
      "逾\n",
      "執照\n",
      "中指\n",
      "墊\n",
      "火車\n",
      "錶\n",
      "傳訊\n",
      "鼓勵\n",
      "股價\n",
      "德清\n",
      "尚未\n",
      "全\n",
      "至\n",
      "理解\n",
      "聊天\n",
      "空\n",
      "更加\n",
      "ま\n",
      "亂\n",
      "大漲\n",
      "今晚\n",
      "把\n",
      "必要\n",
      "後\n",
      "月底\n",
      "向\n",
      "兒童\n",
      "機構\n",
      "沖\n",
      "世博\n",
      "；\n",
      "昨日\n",
      "綠\n",
      "及其\n",
      "語音\n",
      "中\n",
      "大幅\n",
      "並非\n",
      "新\n",
      "首相\n",
      "大都\n",
      "傳統\n",
      "形象\n",
      "職業\n",
      "開設\n",
      "障礙\n",
      "分區\n",
      "吃喝\n",
      "並在\n",
      "倫敦\n",
      "測試\n",
      "人才\n",
      "講話\n",
      "即使\n",
      "ど\n",
      "放棄\n",
      "吵架\n",
      "腳\n",
      "了解\n",
      "上周\n",
      "只\n",
      "韓元\n",
      "小貓\n",
      "一\n",
      "犯案\n",
      "掉\n",
      "才能\n",
      "共有\n",
      "合理\n",
      "搜\n",
      "昨晚\n",
      "大貓\n",
      "成為\n",
      "爆炸\n",
      "反而\n",
      "優\n",
      "在\n",
      "已有\n",
      "滿意\n",
      "困難\n",
      "4\n",
      "紐約\n",
      "被告\n",
      "今日\n",
      "隻\n",
      "協調\n",
      "黑色\n",
      "不管\n",
      "水痘\n",
      "輕人\n",
      "College\n",
      "大賽\n",
      "監\n",
      "保送\n",
      "達\n",
      "強調\n",
      "攝影\n",
      "對方\n",
      "器材\n",
      "美元\n",
      "自救\n",
      "情形\n",
      "甚麼\n",
      "總理\n",
      "時許\n",
      "Lamigo\n",
      "投\n",
      "另行\n",
      "鍾\n",
      "阿富汗\n",
      "按\n",
      "抵達\n",
      "22\n",
      "競選\n",
      "南韓\n",
      "涉嫌\n",
      "清楚\n",
      "攝\n",
      "酒駕\n",
      "外面\n",
      "發動\n",
      "因此\n",
      "吧\n",
      "至於\n",
      "一開始\n",
      "朱\n",
      "貼文\n",
      "否則\n",
      "創辦\n",
      "徹底\n",
      "球隊\n",
      "兩人\n",
      "復\n",
      "英九\n",
      "嚴重\n",
      "當場\n",
      "_\n",
      "瑞典\n",
      "廠\n",
      "詢時\n",
      "of\n",
      "東\n",
      "市長\n",
      "美金\n",
      "勞工\n",
      "上面\n",
      "現\n",
      "建置\n",
      "當然\n",
      "文\n",
      "機器\n",
      "調整\n",
      "含量\n",
      "手中\n",
      "打破\n",
      "投稿\n",
      "創下\n",
      "雜誌\n",
      "消費者\n",
      "社團\n",
      "宣\n",
      "手上\n",
      "）\n",
      "知識\n",
      "帶\n",
      "一樣\n",
      "外電\n",
      "限量\n",
      "破壞\n",
      "記載\n",
      "拉\n",
      "坐\n",
      "症狀\n",
      "原則\n",
      "with\n",
      "法院\n",
      "爆出\n",
      "な\n",
      "常常\n",
      "入境\n",
      "台股\n",
      "未見\n",
      "學\n",
      "值\n",
      "空間\n",
      "來源\n",
      "通過\n",
      "航空\n",
      "日本\n",
      "發現\n",
      "旁邊\n",
      "對\n",
      "評估\n",
      "協議\n",
      "an\n",
      "as\n",
      "逕\n",
      "差距\n",
      "辯論\n",
      "感\n",
      "先發\n",
      "證據\n",
      "作為\n",
      "撲\n",
      "5\n",
      "顯示\n",
      "登上\n",
      "從\n",
      "文化\n",
      "一旁\n",
      "義\n",
      "軍\n",
      "新高\n",
      "美食\n",
      "技術\n",
      "每\n",
      "他\n",
      "最高\n",
      "2008\n",
      "包含\n",
      "季\n",
      "香港\n",
      "方式\n",
      "王溢正\n",
      "新竹\n",
      "反應\n",
      "指控\n",
      "中華民國\n",
      "安妮\n",
      "符合\n",
      "抱怨\n",
      "表情\n",
      "汽車\n",
      "退\n",
      "郭美瑜\n",
      "行車紀\n",
      "性侵\n",
      "凌晨\n",
      "既然\n",
      "且\n",
      "夥伴\n",
      "停下\n",
      "世代\n",
      "系列\n",
      "tw\n",
      "輪\n",
      "張女\n",
      "to\n",
      "議員\n",
      "縣府\n",
      "職災\n",
      "分享\n",
      "一稿\n",
      "西門\n",
      "拍照\n",
      "輔選\n",
      "距離\n",
      "現場\n",
      "員會\n",
      "15\n",
      "細節\n",
      "有人\n",
      "狗\n",
      "粉絲頁\n",
      "飛\n",
      "負責人\n",
      "人民\n",
      "知名\n",
      "憑\n",
      "亞投行\n",
      "做到\n",
      "年\n",
      "核心\n",
      "摸\n",
      "勇士\n",
      "認領\n",
      "滿\n",
      "邀\n",
      "球\n",
      "的\n",
      "每天\n",
      "不敢\n",
      "自家\n",
      "外\n",
      "車子\n",
      "張姓\n",
      "輕\n",
      "價值\n",
      "執政\n",
      "違反\n",
      "當時\n",
      "：\n",
      "介入\n",
      "縣市\n",
      "群組\n",
      "童星\n",
      "具有\n",
      "單\n",
      "失\n",
      "最新\n",
      "戒指\n",
      "張\n",
      "道歉\n",
      "死者\n",
      "改\n",
      "衝擊\n",
      "主題\n",
      "5000\n",
      "主場\n",
      "留言\n",
      "人性\n",
      "流行\n",
      "內容\n",
      "◎\n",
      "加強\n",
      "準備\n",
      "坦承\n",
      "停車\n",
      "紅樓\n",
      "3000\n",
      "小熊\n",
      "po\n",
      "公督盟\n",
      "に\n",
      "或許\n",
      "米\n",
      "印度\n",
      "辛苦\n",
      "劇本\n",
      "被\n",
      "呂志明\n",
      "控制\n",
      "未\n",
      "有意\n",
      "其實\n",
      "自稱\n",
      "出發\n",
      "判決\n",
      "一般\n",
      "騷擾\n",
      "鎖定\n",
      "人事\n",
      "藍綠\n",
      "股\n",
      "制度\n",
      "全程\n",
      "體驗\n",
      "增長\n",
      "周刊\n",
      "爭議\n",
      "6\n",
      "警消\n",
      "立法院\n",
      "便\n",
      "甚至\n",
      "家裡\n",
      "彰化\n",
      "妨害\n",
      "我覺\n",
      "拍\n",
      "負責\n",
      "風險\n",
      "昨天\n",
      "關心\n",
      "影片\n",
      "單車\n",
      "胡\n",
      "火勢\n",
      "環保署\n",
      "澳洲\n",
      "認養\n",
      "我要\n",
      "計畫\n",
      "充滿\n",
      "官方\n",
      "即可\n",
      "法案\n",
      "失焦\n",
      "數\n",
      "公尺\n",
      "比較\n",
      "領域\n",
      "、\n",
      "因素\n",
      "機關\n",
      "女生\n",
      "中心\n",
      "血\n",
      "出門\n",
      "罹患\n",
      "伊朗\n",
      "皮包\n",
      "重新\n",
      "北\n",
      "身體\n",
      "一定\n",
      "有所\n",
      "越來\n",
      "身高\n",
      "調查\n",
      "？\n",
      "開發\n",
      "候選人\n",
      "理由\n",
      "醫藥\n",
      "執行長\n",
      "報\n",
      "台幣\n",
      "る\n",
      "多數\n",
      "局長\n",
      "協商\n",
      "興趣\n",
      "措施\n",
      "面\n",
      "所有\n",
      "才\n",
      "民眾\n",
      "結民報\n",
      "本內容\n",
      "黏\n",
      "南\n",
      "時間\n",
      "下方\n",
      "收據\n",
      "綜合\n",
      "a\n",
      "獲報\n",
      "這裡\n",
      "降低\n",
      "Brennan\n",
      ")\n",
      "比例\n",
      "使\n",
      "刀\n",
      "檢警\n",
      "路上\n",
      "鬆\n",
      "而\n",
      "此次\n",
      "趕\n",
      "團結\n",
      "妝\n",
      "24\n",
      "尋求\n",
      "26\n",
      "27\n",
      "20\n",
      "21\n",
      "事情\n",
      "23\n",
      "這次\n",
      "28\n",
      "29\n",
      "色彩\n",
      "根據\n",
      "廣播電\n",
      "只有\n",
      "一場\n",
      "這款\n",
      "不到\n",
      "只會\n",
      "基本\n",
      "盃\n",
      "瑄\n",
      "動作\n",
      "ˋ\n",
      "李\n",
      "還\n",
      "車禍\n",
      "內\n",
      "幫忙\n",
      "話題\n",
      "～\n",
      "天天\n",
      "支持者\n",
      "日\n",
      "衣服\n",
      "告訴\n",
      "型\n",
      "印尼\n",
      "不願\n",
      "鐵頭\n",
      "早\n",
      "廳\n",
      "出生\n",
      "呂忠吉\n",
      "警車\n",
      "批評\n",
      "比賽\n",
      "角色\n",
      "監督\n",
      "深夜\n",
      "高通\n",
      "能夠\n",
      "校方\n",
      "原來\n",
      "人物\n",
      "電力\n",
      "每周\n",
      "病患\n",
      "海外\n",
      "談判\n",
      "總裁\n",
      "7\n",
      "登錄\n",
      "順利\n",
      "註\n",
      "這段\n",
      "擔心\n",
      "男性\n",
      "狀\n",
      "慢慢\n",
      "手\n",
      "緊\n",
      "背\n",
      "為止\n",
      "大選\n",
      "完整\n",
      "妻子\n",
      "し\n",
      "取\n",
      "到案\n",
      "設備\n",
      "高達\n",
      "臉書粉\n",
      "自身\n",
      "不會\n",
      "曾經\n",
      "六\n",
      "平\n",
      "對外\n",
      "費率\n",
      "今天\n",
      "書\n",
      "好像\n",
      "集中\n",
      "報道\n",
      "。\n",
      "房價\n",
      "菲律賓\n",
      "（\n",
      "逐漸\n",
      "明\n",
      "照顧\n",
      "只是\n",
      "平時\n",
      "工具\n",
      "存在\n",
      "\"\n",
      "議長\n",
      "這樣\n",
      "一天\n",
      "焦點\n",
      "砲\n",
      "公平\n",
      "戶\n",
      "提告\n",
      "區\n",
      "公布\n",
      "呂女\n",
      "感染\n",
      "記錄\n",
      "山區\n",
      "主管\n",
      "三越\n",
      "法\n",
      "便利\n",
      "花粉\n",
      "搭配\n",
      "繞\n",
      "苡\n",
      "恢\n",
      "叫\n",
      "認識\n",
      "高出\n",
      "進入\n",
      "採取\n",
      "大量\n",
      "製\n",
      "提名\n",
      "可見\n",
      "託\n",
      "展示\n",
      "很\n",
      "事表達\n",
      "訓練\n",
      "樑\n",
      "忘記\n",
      "低階核\n",
      "如果\n",
      "驚\n",
      "TEDxTaipei\n",
      "現實\n",
      "猶太人\n",
      "數量\n",
      "上訴\n",
      "說法\n",
      "資訊\n",
      "稱\n",
      "店家\n",
      "雲\n",
      "檢出\n",
      "懷疑\n",
      "葉國吏\n",
      "錄音\n",
      "不夠\n",
      "拒絕\n",
      "不是\n",
      "等等\n",
      "布料\n",
      "棒\n",
      "●\n",
      "70\n",
      "The\n",
      "影像\n",
      "點\n",
      "Google\n",
      "65\n",
      "帶來\n",
      "失去\n",
      "不明\n",
      "遭\n",
      "公庫\n",
      "檢察官\n",
      "打擊\n",
      "飲\n",
      "死\n",
      "擺\n",
      "堅稱\n",
      "認定\n",
      "整合\n",
      "藝術\n",
      "吳春台\n",
      "審議\n",
      "安打\n",
      "族群\n",
      "頭部\n",
      "整理\n",
      "當天\n",
      "仍然\n",
      "並以\n",
      "伊拉克\n",
      "國內\n",
      "合一\n",
      "發表\n",
      "患\n",
      "機會\n",
      "局部\n",
      "強烈\n",
      "南海\n",
      "徵召\n",
      "污染\n",
      "提升\n",
      "收\n",
      "退休\n",
      "加拿大\n",
      "發行\n",
      "挺\n",
      "大隊\n",
      "調\n",
      "如此\n",
      "騎士\n",
      "央行\n",
      "學校\n",
      "高雄\n",
      "跑\n",
      "就\n",
      "將會\n",
      "伊拉\n",
      "權益\n",
      "超級\n",
      "感情\n",
      "費用\n",
      "時候\n",
      "竟\n",
      "大雨\n",
      "巴黎\n",
      "處理\n",
      "實驗\n",
      "執行\n",
      "德國\n",
      "抓\n",
      "一年\n",
      "夫妻\n",
      "小姐\n",
      "追\n",
      "檢討\n",
      "這起\n",
      "富邦\n",
      "財產\n",
      "包\n",
      "設計師\n",
      "部落\n",
      "政黨\n",
      "中立\n",
      "出口\n",
      "出現\n",
      "檔\n",
      "組織\n",
      "2.0\n",
      "大阪\n",
      "蔡\n",
      "自行\n",
      "#\n",
      "一度\n",
      "片\n",
      "很快\n",
      "喝\n",
      "含\n",
      "不慎\n",
      "林姓\n",
      "提及\n",
      "生氣\n",
      "告發\n",
      "留下\n",
      "表現\n",
      "今天下午\n",
      "洪秀柱\n",
      "絲團\n",
      "規模\n",
      "公開\n",
      "立即\n",
      "言論\n",
      "國家\n",
      "出去\n",
      "新增\n",
      "一點\n",
      "歧視\n",
      "教師\n",
      "並且\n",
      "整體\n",
      "可以\n",
      "上海\n",
      "副\n",
      "玩\n",
      "公務員\n",
      "想像\n",
      "開放\n",
      "參加\n",
      "水準\n",
      "申請\n",
      "不想\n",
      "不能\n",
      "儘\n",
      "保證\n",
      "成本\n",
      "壓力\n",
      "移民\n",
      "出版\n",
      "考慮\n",
      "聲音\n",
      "回到\n",
      "訪\n",
      "行政院\n",
      "妳\n",
      "客人\n",
      "一張\n",
      "全會\n",
      "大型\n",
      "列為\n",
      "提到\n",
      "法國\n",
      "合作\n",
      "屆\n",
      "倍\n",
      "參觀\n",
      "新聞\n",
      "動物\n",
      "監視器\n",
      "工地\n",
      "屏東\n",
      "若\n",
      "出刊\n",
      "黨\n",
      "可\n",
      "の\n",
      "罪嫌\n",
      "停止\n",
      "吸毒\n",
      "管線\n",
      "老師\n",
      "很多\n",
      "青少年\n",
      "當下\n",
      "創作\n",
      "採用\n",
      "裸照\n",
      "本月\n",
      "36\n",
      "萬餘元\n",
      "保母\n",
      "差\n",
      "RCA\n",
      "團體\n",
      "增\n",
      "生活\n",
      "辦法\n",
      "今年\n",
      "面對\n",
      "說\n",
      "節目\n",
      "錯\n",
      "那些\n",
      "科技\n",
      "那亞\n",
      "完成\n",
      "難道\n",
      "站台\n",
      "私人\n",
      "導演\n",
      "很大\n",
      "量刑\n",
      "魏揚\n",
      "只能\n",
      "元\n",
      "威脅\n",
      "女兒\n",
      "安排\n",
      "網站\n",
      "選人\n",
      "似乎\n",
      "股市\n",
      "比\n",
      "咪咪\n",
      "す\n",
      "壞\n",
      "周四\n",
      "那麼\n",
      "八卦\n",
      "努力\n",
      "女\n",
      "幾個\n",
      "美方\n",
      "業者\n",
      "要求\n",
      "安\n",
      "圖\n",
      "開出\n",
      "周邊\n",
      "特定\n",
      "場\n",
      "正面\n",
      "待\n",
      "款\n",
      "...\n",
      "獲得\n",
      "領導\n",
      "い\n",
      "圖文\n",
      "媽媽\n",
      "Cayenne\n",
      "藍\n",
      "呼吸\n",
      "近\n",
      "協助\n",
      "同學\n",
      "女孩\n",
      "敢\n",
      "d\n",
      "機制\n",
      "房間\n",
      "廢料\n",
      "國外\n",
      "發言\n",
      "以外\n",
      "藉此\n",
      "討論\n",
      "但是\n",
      "北市\n",
      "將\n",
      "提出\n",
      "休\n",
      "有效\n",
      "樓\n",
      "動物園\n",
      "槍手\n",
      "屠殺\n",
      "計劃\n",
      "愈來\n",
      "觀察\n",
      "原能會\n",
      "傳出\n",
      "辦\n",
      "被迫\n",
      "咬\n",
      "財報\n",
      "網\n",
      "近來\n",
      "貼\n",
      "be\n",
      "公宅\n",
      "單位\n",
      "300\n",
      "自己\n",
      "年間\n",
      "佑\n",
      "強制\n",
      "from\n",
      "坐在\n",
      "by\n",
      "主席\n",
      "提醒\n",
      "公室\n",
      "TVBS\n",
      "小型\n",
      "營收\n",
      "&\n",
      "大陸\n",
      "處置\n",
      "消息\n",
      "開車\n",
      "增加\n",
      "為何\n",
      "臺\n",
      "需\n",
      "億\n",
      "而是\n",
      "至少\n",
      "期間\n",
      "資料\n",
      "讓\n",
      "投入\n",
      "統計\n",
      "日起\n",
      "報告\n",
      "等待\n",
      "繳費\n",
      "活動\n",
      "並\n",
      "工商\n",
      "行車\n",
      "治療\n",
      "所致\n",
      "抹\n",
      ":\n",
      "購買\n",
      "參\n",
      "呂\n",
      "人士\n",
      "地點\n",
      "持續\n",
      "女友\n",
      "男生\n",
      "這也\n",
      "處分\n",
      "做\n",
      "報名\n",
      "香鯨\n",
      "外交\n",
      "任何\n",
      "上任\n",
      "中央\n",
      "....\n",
      "之前\n",
      "罰\n",
      "水質\n",
      "年齡\n",
      "賽\n",
      "予以\n",
      "立刻\n",
      "教授\n",
      "產業\n",
      "全新\n",
      "珍妮\n",
      "家中\n",
      "一經\n",
      "這些\n",
      "沒\n",
      "市民\n",
      "英國\n",
      "嘛\n",
      "%\n",
      "夫婦\n",
      "事故\n",
      "亦\n",
      "經驗\n",
      "老公\n",
      "孔子\n",
      "交流\n",
      "萬戶\n",
      "整個\n",
      "路段\n",
      "球迷\n",
      "全文\n",
      "創新\n",
      "挑戰\n",
      "主張\n",
      "嫌犯\n",
      "支持\n",
      "升高\n",
      "用水\n",
      "雖\n",
      "機率\n",
      "法令\n",
      "每年\n",
      "媒體\n",
      "張博亭\n",
      "閱讀\n",
      "每月\n",
      "強化\n",
      "2000\n",
      "捕手\n",
      "at\n",
      "菸\n",
      "結合\n",
      "這件\n",
      "羅\n",
      "一切\n",
      "去年\n",
      "合\n",
      "乾淨\n",
      "保時\n",
      "gl\n",
      "求婚\n",
      "貢獻\n",
      "開心\n",
      "建案\n",
      "所以\n",
      "人們\n",
      "看此\n",
      "王姓\n",
      "亞馬遜\n",
      "雖然\n",
      "整形\n",
      "希望\n",
      "放\n",
      "一片\n",
      "娃娃\n",
      "呈\n",
      "害怕\n",
      "聯盟\n",
      "P\n",
      "改權\n",
      "facebook\n",
      "不足\n",
      "時尚\n",
      "環境\n",
      "皇家\n",
      "賣\n",
      "穩定\n",
      "蹤\n",
      "工程\n",
      "季後賽\n",
      "看看\n",
      "醫學\n",
      "北部\n",
      "獲\n",
      "?\n",
      "替\n",
      "％\n",
      "十分\n",
      "搓\n",
      "享受\n",
      "原文\n",
      "審查\n",
      "人家\n",
      "3D\n",
      "方案\n",
      "實際\n",
      "房子\n",
      "孩子\n",
      "趕緊\n",
      ";\n",
      "國民黨\n",
      "年代\n",
      "利率\n",
      "戰爭\n",
      "myVideo\n",
      "則\n",
      "已經\n",
      "捐款\n",
      "進站\n",
      "插\n",
      "本報\n",
      "認為\n",
      "39\n",
      "38\n",
      "聲稱\n",
      "33\n",
      "32\n",
      "31\n",
      "30\n",
      "明智\n",
      "35\n",
      "34\n",
      "共\n",
      "名媛\n",
      "防部\n",
      "分析\n",
      "大家\n",
      "超過\n",
      "率\n",
      "警方\n",
      "見\n",
      "相信\n",
      "公告\n",
      "實在\n",
      "免費\n",
      "選手\n",
      "呈現\n",
      "犯罪\n",
      "…\n",
      "台電\n",
      "第\n",
      "能力\n",
      "照片\n",
      "交換\n",
      "負擔\n",
      "得到\n",
      "攻擊\n",
      "給\n",
      "出席\n",
      "展現\n",
      "身分\n",
      "廣播\n",
      "う\n",
      "匯率\n",
      "中國\n",
      "目標\n",
      "愛情\n",
      "8\n",
      "選戰\n",
      "勝\n",
      "費\n",
      "裡\n",
      "們\n",
      "選舉\n",
      "以\n",
      "說明\n",
      "知道\n",
      "網路\n",
      "繫\n",
      "納入\n",
      "品質\n",
      "Police\n",
      "登場\n",
      "隊\n",
      "生意\n",
      "his\n",
      "生產\n",
      "庚\n",
      "愛\n",
      "真\n",
      "少女\n",
      "每個\n",
      "四年\n",
      "結婚\n",
      "專家\n",
      "下跌\n",
      "期\n",
      "巴基斯坦\n",
      "危險\n",
      "疫情\n",
      "團隊\n",
      "出庭\n",
      "含有\n",
      "核武\n",
      "模擬\n",
      "住\n",
      "大學\n",
      "Q\n",
      "院長\n",
      "處\n",
      "字\n",
      "年來\n",
      "法務部\n",
      "設定\n",
      "疾病\n",
      "老鼠\n",
      "改變\n",
      "高雄市\n",
      "一句\n",
      "自動\n",
      "門票\n",
      "本次\n",
      "寬\n",
      "更是\n",
      "一台\n",
      "運動\n",
      "國際\n",
      "真正\n",
      "威仁\n",
      "北水處\n",
      "小孩\n",
      "一個\n",
      "自然\n",
      "特別\n",
      "四川\n",
      "成\n",
      "支付\n",
      "每周五\n",
      "政策\n",
      "級\n",
      "孩童\n",
      "形容\n",
      "優勢\n",
      "消費\n",
      "主動\n",
      "昨\n",
      "onlineopinions\n",
      "粉絲\n",
      "然而\n",
      "敲出\n",
      "到\n",
      "廚房\n",
      "喵\n",
      "2015\n",
      "2014\n",
      "2016\n",
      "2011\n",
      "2010\n",
      "2013\n",
      "2012\n",
      "建設\n",
      "黃\n",
      "效果\n",
      "請\n",
      "啊\n",
      "最低\n",
      "行\n",
      "轎\n",
      "大部分\n",
      "理想\n",
      "公園\n",
      "iPhone\n",
      "撤銷\n",
      "新低\n",
      "停\n",
      "非\n",
      "人\n",
      "水事業\n",
      "硬\n",
      "更新\n",
      "最佳\n",
      "全案\n",
      "作者\n",
      "三重\n",
      "能量\n",
      "觀眾\n",
      "老婦\n",
      "造勢\n",
      "今天上午\n",
      "聯名\n",
      "如\n",
      "聯合\n",
      "多久\n",
      "喊\n",
      "解決\n",
      "頗\n",
      "販售\n",
      "獲獎\n",
      "收費\n",
      "將發布\n",
      "瞞\n",
      "次\n",
      "不但\n",
      "'\n",
      "等到\n",
      "雇主\n",
      "坦言\n",
      "金門\n",
      "營業\n",
      "人生\n",
      "太陽能\n",
      "現狀\n",
      "車內\n",
      "處長\n",
      "或者\n",
      "假\n",
      "如今\n",
      "飲料\n",
      "公社\n",
      "地區\n",
      "可惜\n",
      "損失\n",
      "彙\n",
      "鉛\n",
      "線\n",
      "力量\n",
      "哪裡\n",
      "蜜粉\n",
      "功能\n",
      "便宜\n",
      "另外\n",
      "一部\n",
      "具體\n",
      "祭\n",
      "可能\n",
      "深入\n",
      "追求\n",
      "生態\n",
      "繼續\n",
      "將在\n",
      "領先\n",
      "頂\n",
      "約\n",
      "分\n",
      "爆料\n",
      "衝突\n",
      "兩次\n",
      "轉型\n",
      "就是\n",
      "不滿\n",
      "建議\n",
      "晚間\n",
      "確認\n",
      "而言\n",
      "出局\n",
      "鶯\n",
      "來稿\n",
      "進\n",
      "登革熱\n",
      "病毒\n",
      "萬人\n",
      "商業\n",
      "心\n",
      "組\n",
      "好好\n",
      "邀請\n",
      "告\n",
      "人員\n",
      "滑板\n",
      "如何\n",
      "台南市\n",
      "胡男\n",
      "外交部\n",
      "洪辦\n",
      "導致\n",
      "Apple\n",
      "樁\n",
      "電子\n",
      "字樣\n",
      "理念\n",
      "寫\n",
      "檢舉\n",
      "早上\n",
      "婚姻\n",
      "不如\n",
      "能\n",
      "are\n",
      "教育部\n",
      "柯文\n",
      "啟動\n",
      "電影\n",
      "事業\n",
      "眼睛\n",
      "我\n",
      "阻止\n",
      "have\n",
      "住家\n",
      "生\n",
      "內部\n",
      "還在\n",
      "單純\n",
      "馬\n",
      "累\n",
      "選情\n",
      "絕對\n",
      "千元\n",
      "鐵\n",
      "肛交\n",
      "truck\n",
      "=\n",
      "當事人\n",
      "實驗室\n",
      "身邊\n",
      "起火\n",
      "00\n",
      "億元\n",
      "提\n",
      "南投\n",
      "忘\n",
      "回\n",
      "互動\n",
      "下來\n",
      "刊登\n",
      "概念\n",
      "避免\n",
      "版本\n",
      "fjs\n",
      "身為\n",
      "賺\n",
      "聽\n",
      "獲利\n",
      "蘋果\n",
      "綁\n",
      "胸部\n",
      "更換\n",
      "資源\n",
      "有期徒刑\n",
      "模特\n",
      "試探\n",
      "事件\n",
      "店\n",
      "趨勢\n",
      "友人\n",
      "氣象局\n",
      "願意\n",
      "可是\n",
      "經營\n",
      "新聞區\n",
      "並沒有\n",
      "困擾\n",
      "本身\n",
      "應用\n",
      "傳\n",
      "之外\n",
      "公共\n",
      "給予\n",
      "報導\n",
      "方法\n",
      "消防局\n",
      "敘利亞\n",
      "財政部\n",
      "局\n",
      "水資源\n",
      "臨\n",
      "攀升\n",
      "商\n",
      "半\n",
      "回答\n",
      "下載\n",
      "人口\n",
      "短\n",
      "點半\n",
      "不快\n",
      "針對\n",
      "願\n",
      "總是\n",
      "男人\n",
      "糾紛\n",
      "不要\n",
      "給付\n",
      "認同\n",
      "蔣翁\n",
      "不見\n",
      "基金\n",
      "亮相\n",
      "時報\n",
      "首先\n",
      "操作\n",
      "接受\n",
      "拿\n",
      "燈\n",
      "經典\n",
      "台南\n",
      "老婆\n",
      "參考\n",
      "此時\n",
      "公里\n",
      "鎊\n",
      "工廠\n",
      "關切\n",
      "非常\n",
      "專頁\n",
      "就算\n",
      "以及\n",
      "別墅\n",
      "您\n",
      "超越\n",
      "回家\n",
      "由\n",
      "病人\n",
      "殺\n",
      "PO\n",
      "房\n",
      "不必\n",
      "維持\n",
      "台北\n",
      "嗆\n",
      "及\n",
      "位\n",
      "救\n",
      "將來\n",
      "S\n",
      "近日\n",
      "行政\n",
      "期限\n",
      "檢方\n",
      "接獲\n",
      "得知\n",
      "店員\n",
      "恐怕\n",
      "或是\n",
      "依照\n",
      "奇摩\n",
      "一直\n",
      "擴大\n",
      "判刑\n",
      "進黨\n",
      "交易\n",
      "本來\n",
      "妃\n",
      "市\n",
      "ら\n",
      "國聯冠\n",
      "中午\n",
      "可望\n",
      "公道\n",
      "得\n",
      "材質\n",
      "檢驗\n",
      "期待\n",
      "共同\n",
      "原\n",
      "以前\n",
      "踢\n",
      "秘書長\n",
      "感覺\n",
      "錄用\n",
      "設置\n",
      "Yahoo\n",
      "偶像\n",
      "原本\n",
      "拿下\n",
      "包括\n",
      "人力\n",
      ">\n",
      "日報\n",
      "有些\n",
      "り\n",
      "先\n",
      "救出\n",
      "仍\n",
      "召開\n",
      "周年\n",
      "看到\n",
      "保持\n",
      "結構\n",
      "婚\n",
      "名人\n",
      "核電站\n",
      "此\n",
      "旅遊\n",
      "王粉\n",
      "逼\n",
      "店長\n",
      "惡化\n",
      "生日\n",
      "~\n",
      "暴力\n",
      "病例\n",
      "引起\n",
      "立場\n",
      "不\n",
      "電話\n",
      "溫泉\n",
      "地球\n",
      "士林\n",
      "露出\n",
      "傷勢\n",
      "罪判\n",
      "接到\n",
      "粉絲團\n",
      "重要\n",
      "晶片\n",
      "玩笑\n",
      "提高\n",
      "郵報\n",
      "指出\n",
      "符號\n",
      "重視\n",
      "兩位\n",
      "注意\n",
      "規劃\n",
      "恐慌\n",
      "助理\n",
      "所謂\n",
      "水號\n",
      "再度\n",
      "http\n",
      "寄\n",
      "均\n",
      "論壇\n",
      "衛生\n",
      "情況\n",
      "野餐\n",
      "嘴\n",
      "泰國\n",
      "並未\n",
      "低\n",
      "預算\n",
      "想到\n",
      "徵稿\n",
      "關鍵\n",
      "電線\n",
      "詢問\n",
      "刺激\n",
      "受訪\n",
      "直到\n",
      "希拉\n",
      "預定\n",
      "酒\n",
      "創\n",
      "三振\n",
      "47\n",
      "簡單\n",
      "房市\n",
      "表示\n",
      "呼籲\n",
      "飛安會\n",
      "事\n",
      "運用\n",
      "自由\n",
      "一同\n",
      "查詢\n",
      "施工\n",
      "進行\n",
      "死亡\n",
      "主因\n",
      "—\n",
      "偵組\n",
      "尤其\n",
      "多\n",
      "二人\n",
      "區域\n",
      "議題\n",
      "開始\n",
      "圓仔湯\n",
      "YouTube\n",
      "考試\n",
      "近期\n",
      "加重\n",
      "承認\n",
      "買\n",
      "夜景\n",
      "又彥\n",
      "承諾\n",
      "屋內\n",
      "根本\n",
      "版\n",
      "尊重\n",
      "告知\n",
      "『\n",
      "瑜\n",
      "政府\n",
      "洛杉磯\n",
      "錄器\n",
      "屬\n",
      "飯\n",
      "賠償\n",
      "餐廳\n",
      "立委\n",
      "有趣\n",
      "女童\n",
      "！\n",
      "檢查\n",
      "林\n",
      "陽耀勳\n",
      "加入\n",
      "下\n",
      "熊\n",
      "參與\n",
      "娛樂\n",
      "地震\n",
      "證明\n",
      "介紹\n",
      "唯不付\n",
      "小心\n",
      "不得\n",
      "陪同\n",
      "發票\n",
      "明星\n",
      "藝術家\n",
      "情侶\n",
      "公司\n",
      "彭麗媛\n",
      "槍\n",
      "各種\n",
      "組合\n",
      "之\n",
      "保護\n",
      "輸入\n",
      "打\n",
      "調閱\n",
      "高\n",
      "た\n",
      "教育\n",
      "各自\n",
      "林緯\n",
      "幫\n",
      "自\n",
      "退件\n",
      "緊急\n",
      "希臘\n",
      "定位\n",
      "它\n",
      "聽到\n",
      "簽署\n",
      "題目\n",
      "例\n",
      "《\n",
      "引述\n",
      "以色列\n",
      "投票\n",
      "音樂\n",
      "歡迎\n",
      "快速\n",
      "健保\n",
      "項目\n",
      "這是\n",
      "幫助\n",
      "柯建銘\n",
      "看過\n",
      "1989\n",
      "地\n",
      "而且\n",
      "不論\n",
      "展開\n",
      "交通部\n",
      "帶領\n",
      "速度\n",
      "喇叭\n",
      "社會\n",
      "闆\n",
      "公分\n",
      "系統\n",
      "影展\n",
      "症\n",
      "臉書\n",
      "選擇\n",
      "搶\n",
      "競爭\n",
      "現在\n",
      "昨在\n",
      "旅客\n",
      "方便\n",
      "站\n",
      "he\n",
      "台灣人\n",
      "你\n",
      "協定\n",
      "博士\n",
      "快\n",
      "反光\n",
      "高工\n",
      "右\n",
      "決定\n",
      "抗議\n",
      "認真\n",
      "其他\n",
      "急救\n",
      "辯稱\n",
      "找\n",
      "往\n",
      "感謝\n",
      "項\n",
      "不佳\n",
      "皆\n",
      "路口\n",
      "芳苑\n",
      "管制\n",
      "母親\n",
      "一間\n",
      "範圍\n",
      "怕\n",
      "隔天\n",
      "發揮\n",
      "證會\n",
      "80\n",
      "致死\n",
      "48\n",
      "49\n",
      "46\n",
      "推\n",
      "44\n",
      "45\n",
      "42\n",
      "即\n",
      "40\n",
      "41\n",
      "台灣\n",
      "大約\n",
      "房屋\n",
      "身上\n",
      "一名\n",
      "他人\n",
      "白粉\n",
      "於\n",
      "多歲\n",
      "李男\n",
      "水利\n",
      "又\n",
      "女子\n",
      "信心\n",
      "銀行\n",
      "新北市\n",
      "以為\n",
      "一位\n",
      "關\n",
      "董事\n",
      "正\n",
      "性\n",
      "造型\n",
      "時代\n",
      "多次\n",
      "髮\n",
      "地院\n",
      "我會\n",
      "哲\n",
      "園區\n",
      "腿\n",
      "起\n",
      "暨\n",
      "租金\n",
      "加速\n",
      "遊客\n"
     ]
    }
   ],
   "source": [
    "for ele in model.vocab:\n",
    "    print ele"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 470,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "水事業 0.985005259514\n",
      "自來 0.984110593796\n",
      "水號 0.982387363911\n",
      "水管 0.981813788414\n",
      "鉛 0.981554508209\n",
      "水質 0.978017866611\n",
      "北水處 0.976608633995\n",
      "用水 0.974708735943\n",
      "更換 0.968997180462\n",
      "路段 0.963771760464\n"
     ]
    }
   ],
   "source": [
    "for w in model.most_similar('鉛管'.decode('utf-8')):\n",
    "    print w[0], w[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 471,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.save('news_wv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 472,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model2 = word2vec.Word2Vec.load('news_wv')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 473,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "德清 0.99296349287\n",
      "八仙 0.987828493118\n",
      "門票 0.985619783401\n",
      "林緯 0.967615008354\n",
      "自救 0.964752078056\n",
      "局長 0.959368765354\n",
      "錄音 0.956847786903\n",
      "一事 0.956403493881\n",
      "小姐 0.95299577713\n",
      "公務員 0.95046710968\n"
     ]
    }
   ],
   "source": [
    "for w in model.most_similar('呂忠吉'.decode('utf-8')):\n",
    "    print w[0], w[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}