Text Analytics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Common use cases"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Curriculum\n",
    "\n",
    "1. Introduction <br>\n",
    "<br>\n",
    "2. Text feature extraction  / Text vectorization - feature engineering for ML algorithms<br>\n",
    "    a. Frequency Vectors<br>\n",
    "    b. One-hot Encoding<br>\n",
    "    c. Term Frequency - Inverse Document Frequency (TF-IDF)<br>\n",
    "    <br>\n",
    "3. Text clustering by Topic Modeling - abstract core themes from a set of documents<br>\n",
    "    a. LDA<br>\n",
    "    b. LSA<br>\n",
    "    c. NMF<br>\n",
    "    <br>\n",
    "4. Text categorization<br>\n",
    "    a. Linear Classifiers<br>\n",
    "    b. Gaussian Processes - SVM<br>\n",
    "    c. Tree-based classifiers - Random Forest, Gradient Boosting Machines<br>\n",
    "    <br>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Machine learning algorithms operate on a numeric feature space, expecting input as a two-dimensional array where rows are instances and columns are features. **In order to perform machine learning on text, we need to transform our documents into vector representations such that we can apply numeric machine learning.** This process is called feature extraction or more simply, vectorization, and is an essential first step in text analytics."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### What is the Sentiment?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\"The movie Amazing Spiderman is bad\"<br>\n",
    "<br>\n",
    "“I had high hopes and great expectations for the movie dubbed wonderful and exhilarating by critics, but was hugely disappointed.\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Text feature extraction using Bag-of-words approach"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.Term Frequency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "corpus = [\n",
    "        \"Your eligibility depends upon your city, salary, residence type, age, employment status, and credit score. If you have a good credit score, you can bargain more.\",\n",
    "        \"It’s never a good idea to drag a loan for many years or you could end up losing more money.\",\n",
    "        \"It is advisable to think long-term, so think till how long you can keep paying off the loan. It is always a good idea to keep some money aside for emergencies.\",\n",
    "        ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "vectorizer = CountVectorizer()\n",
    "vectors = vectorizer.fit_transform(corpus)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['advisable', 'age', 'always', 'and', 'aside', 'bargain', 'can', 'city', 'could', 'credit', 'depends', 'drag', 'eligibility', 'emergencies', 'employment', 'end', 'for', 'good', 'have', 'how', 'idea', 'if', 'is', 'it', 'keep', 'loan', 'long', 'losing', 'many', 'money', 'more', 'never', 'off', 'or', 'paying', 'residence', 'salary', 'score', 'so', 'some', 'status', 'term', 'the', 'think', 'till', 'to', 'type', 'up', 'upon', 'years', 'you', 'your']\n",
      "52\n"
     ]
    }
   ],
   "source": [
    "print(vectorizer.get_feature_names())\n",
    "\n",
    "print(len(vectorizer.get_feature_names()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 1 0 1 0 1 1 1 0 2 1 0 1 0 1 0 0 1 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1\n",
      "  1 2 0 0 1 0 0 0 0 0 1 0 1 0 2 2]\n",
      " [0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 1 1 1 0 0 1 0 0 1 0 1 0 1 1 1 1 1 0 1 0 0\n",
      "  0 0 0 0 0 0 0 0 0 1 0 1 0 1 1 0]\n",
      " [1 0 1 0 1 0 1 0 0 0 0 0 0 1 0 0 1 1 0 1 1 0 2 2 2 1 2 0 0 1 0 0 1 0 1 0\n",
      "  0 0 1 1 0 1 1 2 1 2 0 0 0 0 1 0]]\n",
      "(3, 52)\n"
     ]
    }
   ],
   "source": [
    "print(vectors.toarray())\n",
    "print(vectors.toarray().shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "cloud_words = ' '\n",
    "\n",
    "for val in corpus:\n",
    " # typecaste each val to string\n",
    "    val = str(val)\n",
    " \n",
    " # split the value\n",
    "    tokens = val.split()\n",
    "     \n",
    "        # Converts each token into lowercase\n",
    "    for i in range(len(tokens)):\n",
    "        tokens[i] = tokens[i].lower()\n",
    "         \n",
    "    for words in tokens:\n",
    "        cloud_words = cloud_words + words + ' '"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\" taggeddocument(<generator object tokenize at 0x0000018390c062b0>, ['d0']) taggeddocument(<generator object tokenize at 0x0000018390c06bf8>, ['d1']) taggeddocument(<generator object tokenize at 0x0000018390c06ca8>, ['d2']) \""
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cloud_words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[TaggedDocument(words=<generator object tokenize at 0x0000018393A8F6D0>, tags=['d0']),\n",
       " TaggedDocument(words=<generator object tokenize at 0x0000018393A8F678>, tags=['d1']),\n",
       " TaggedDocument(words=<generator object tokenize at 0x0000018393A8F728>, tags=['d2'])]"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corpus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from wordcloud import WordCloud, STOPWORDS \n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import pandas as pd \n",
    "\n",
    "corpus = [\n",
    "    \"Your eligibility depends upon your city, salary, residence type, age, employment status, and credit score. If you have a good credit score, you can bargain more.\",\n",
    "    \"It’s never a good idea to drag a loan for many years or you could end up losing more money.\",\n",
    "    \"It is advisable to think long-term, so think till how long you can keep paying off the loan. It is always a good idea to keep some money aside for emergencies.\"\n",
    "]\n",
    "\n",
    "cloud_words = ' '\n",
    "\n",
    "for val in corpus:\n",
    " # typecaste each val to string\n",
    "    val = str(val)\n",
    " \n",
    " # split the value\n",
    "    tokens = val.split()\n",
    "     \n",
    "        # Converts each token into lowercase\n",
    "    for i in range(len(tokens)):\n",
    "        tokens[i] = tokens[i].lower()\n",
    "         \n",
    "    for words in tokens:\n",
    "        cloud_words = cloud_words + words + ' '\n",
    "\n",
    "wordcloud = WordCloud(width = 800, height = 800,background_color ='white',min_font_size = 10).generate(cloud_words)\n",
    "    # print(wordcloud)\n",
    "\n",
    "# plot the WordCloud image                       \n",
    "plt.figure(figsize = (8, 8), facecolor = None)\n",
    "plt.imshow(wordcloud)\n",
    "plt.axis(\"off\")\n",
    "plt.tight_layout(pad = 0)\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Frequency-based encoding methods suffer from the long tail, or Zipfian distribution, that characterizes natural language. As a result, tokens that occur very frequently are orders of magnitude more “significant” than other, less frequent ones. This can have a significant impact on some models (e.g., generalized linear models) that expect normally distributed features."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.One-hot encoding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### A solution to this problem is one-hot encoding, a boolean vector encoding method that marks a particular vector index with a value of true (1) if the token exists in the document and false (0) if it does not."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import Binarizer\n",
    "\n",
    "vectorizer = CountVectorizer()\n",
    "vectors = vectorizer.fit_transform(corpus)\n",
    "\n",
    "onehot = Binarizer()\n",
    "vectors_onehot = onehot.fit_transform(vectors.toarray())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 1 0 1 0 1 1 1 0 1 1 0 1 0 1 0 0 1 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1\n",
      "  1 1 0 0 1 0 0 0 0 0 1 0 1 0 1 1]\n",
      " [0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 1 1 1 0 0 1 0 0 1 0 1 0 1 1 1 1 1 0 1 0 0\n",
      "  0 0 0 0 0 0 0 0 0 1 0 1 0 1 1 0]\n",
      " [1 0 1 0 1 0 1 0 0 0 0 0 0 1 0 0 1 1 0 1 1 0 1 1 1 1 1 0 0 1 0 0 1 0 1 0\n",
      "  0 0 1 1 0 1 1 1 1 1 0 0 0 0 1 0]]\n",
      "(3, 52)\n"
     ]
    }
   ],
   "source": [
    "print(vectors_onehot)\n",
    "\n",
    "print(vectors_onehot.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 1 0 1 0 1 1 1 0 1 1 0 1 0 1 0 0 1 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1\n",
      "  1 1 0 0 1 0 0 0 0 0 1 0 1 0 1 1]\n",
      " [0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 1 1 1 0 0 1 0 0 1 0 1 0 1 1 1 1 1 0 1 0 0\n",
      "  0 0 0 0 0 0 0 0 0 1 0 1 0 1 1 0]\n",
      " [1 0 1 0 1 0 1 0 0 0 0 0 0 1 0 0 1 1 0 1 1 0 1 1 1 1 1 0 0 1 0 0 1 0 1 0\n",
      "  0 0 1 1 0 1 1 1 1 1 0 0 0 0 1 0]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import Binarizer\n",
    "\n",
    "del vectors\n",
    "vectorizer = CountVectorizer(binary=True)\n",
    "vectors = vectorizer.fit_transform(corpus)\n",
    "\n",
    "print(vectors.toarray())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.Term Frequency - Inverse Document Frequency"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A better approach would be to consider the relative frequency or rareness of tokens in the document against their frequency in other documents. <br>\n",
    "<br>\n",
    "TF–IDF encoding normalizes the frequency of tokens in a document with respect to the rest of the corpus. It accentuates terms that are very relevant to a specific document.\n",
    "<br>\n",
    "<br>\n",
    "The central insight is that meaning is most likely encoded in the more rare terms from a document. For example, in a corpus of sports text, tokens such as “umpire,” “base,” and “dugout” appear more frequently in documents that discuss baseball, while other tokens that appear frequently throughout the corpus, like “run,” “score,” and “play,” are less important.<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We interpret the score to mean that the closer the TF–IDF score of a term is to 1, the more informative that term is to that document. The closer the score is to zero, the less informative that term is."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "tfidf = TfidfVectorizer()\n",
    "corpus = tfidf.fit_transform(corpus)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.         0.18601332 0.         0.18601332 0.         0.18601332\n",
      "  0.14146793 0.18601332 0.         0.37202664 0.18601332 0.\n",
      "  0.18601332 0.         0.18601332 0.         0.         0.10986246\n",
      "  0.18601332 0.         0.         0.18601332 0.         0.\n",
      "  0.         0.         0.         0.         0.         0.\n",
      "  0.14146793 0.         0.         0.         0.         0.18601332\n",
      "  0.18601332 0.37202664 0.         0.         0.18601332 0.\n",
      "  0.         0.         0.         0.         0.18601332 0.\n",
      "  0.18601332 0.         0.21972493 0.37202664]\n",
      " [0.         0.         0.         0.         0.         0.\n",
      "  0.         0.         0.26971476 0.         0.         0.26971476\n",
      "  0.         0.         0.         0.26971476 0.20512502 0.15929788\n",
      "  0.         0.         0.20512502 0.         0.         0.20512502\n",
      "  0.         0.20512502 0.         0.26971476 0.26971476 0.20512502\n",
      "  0.20512502 0.26971476 0.         0.26971476 0.         0.\n",
      "  0.         0.         0.         0.         0.         0.\n",
      "  0.         0.         0.         0.20512502 0.         0.26971476\n",
      "  0.         0.26971476 0.15929788 0.        ]\n",
      " [0.16616696 0.         0.16616696 0.         0.16616696 0.\n",
      "  0.12637426 0.         0.         0.         0.         0.\n",
      "  0.         0.16616696 0.         0.         0.12637426 0.09814088\n",
      "  0.         0.16616696 0.12637426 0.         0.33233392 0.25274851\n",
      "  0.33233392 0.12637426 0.33233392 0.         0.         0.12637426\n",
      "  0.         0.         0.16616696 0.         0.16616696 0.\n",
      "  0.         0.         0.16616696 0.16616696 0.         0.16616696\n",
      "  0.16616696 0.33233392 0.16616696 0.25274851 0.         0.\n",
      "  0.         0.         0.09814088 0.        ]]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>42</th>\n",
       "      <th>43</th>\n",
       "      <th>44</th>\n",
       "      <th>45</th>\n",
       "      <th>46</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>50</th>\n",
       "      <th>51</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.186013</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.186013</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.186013</td>\n",
       "      <td>0.141468</td>\n",
       "      <td>0.186013</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.372027</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.186013</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.186013</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.219725</td>\n",
       "      <td>0.372027</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.269715</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.205125</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.269715</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.269715</td>\n",
       "      <td>0.159298</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.166167</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.166167</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.166167</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.126374</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.166167</td>\n",
       "      <td>0.332334</td>\n",
       "      <td>0.166167</td>\n",
       "      <td>0.252749</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.098141</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 52 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         0         1         2         3         4         5         6   \\\n",
       "0  0.000000  0.186013  0.000000  0.186013  0.000000  0.186013  0.141468   \n",
       "1  0.000000  0.000000  0.000000  0.000000  0.000000  0.000000  0.000000   \n",
       "2  0.166167  0.000000  0.166167  0.000000  0.166167  0.000000  0.126374   \n",
       "\n",
       "         7         8         9   ...        42        43        44        45  \\\n",
       "0  0.186013  0.000000  0.372027  ...  0.000000  0.000000  0.000000  0.000000   \n",
       "1  0.000000  0.269715  0.000000  ...  0.000000  0.000000  0.000000  0.205125   \n",
       "2  0.000000  0.000000  0.000000  ...  0.166167  0.332334  0.166167  0.252749   \n",
       "\n",
       "         46        47        48        49        50        51  \n",
       "0  0.186013  0.000000  0.186013  0.000000  0.219725  0.372027  \n",
       "1  0.000000  0.269715  0.000000  0.269715  0.159298  0.000000  \n",
       "2  0.000000  0.000000  0.000000  0.000000  0.098141  0.000000  \n",
       "\n",
       "[3 rows x 52 columns]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(corpus.toarray())\n",
    "pd.DataFrame(corpus.toarray())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4. Distributed representations - word2vec, doc2vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "corpus = [\n",
    "        \"Your eligibility depends upon your city, salary, residence type, age, employment status, and credit score. If you have a good credit score, you can bargain more.\",\n",
    "        \"It’s never a good idea to drag a loan for many years or you could end up losing more money.\",\n",
    "        \"It is advisable to think long-term, so think till how long you can keep paying off the loan. It is always a good idea to keep some money aside for emergencies.\",\n",
    "        ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.03843604  0.01628663  0.05468123 -0.01201195  0.03241907 -0.03380773\n",
      "  0.06233216]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\admin\\venv\\lib\\site-packages\\gensim\\models\\doc2vec.py:580: UserWarning: The parameter `size` is deprecated, will be removed in 4.0.0, use `vector_size` instead.\n",
      "  warnings.warn(\"The parameter `size` is deprecated, will be removed in 4.0.0, use `vector_size` instead.\")\n"
     ]
    }
   ],
   "source": [
    "from gensim.models.doc2vec import TaggedDocument, Doc2Vec\n",
    "\n",
    "\n",
    "corpus = [list(tokenize(doc)) for doc in corpus]\n",
    "corpus = [TaggedDocument(words, ['d{}'.format(idx)]) for idx, words in enumerate(corpus)]\n",
    "\n",
    "model = Doc2Vec(corpus, size=7, min_count=0)\n",
    "print(model.docvecs[0])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.03843604  0.01628663  0.05468123 -0.01201195  0.03241907 -0.03380773\n",
      "  0.06233216]\n",
      "[ 0.03889636 -0.05206067  0.0557959  -0.03954409  0.0697253   0.038394\n",
      "  0.04519154]\n",
      "[-0.03663519 -0.06676246  0.03769557 -0.01243174 -0.00505658  0.00360704\n",
      "  0.02614829]\n"
     ]
    },
    {
     "ename": "KeyError",
     "evalue": "\"tag '3' not seen in training corpus/invalid\"",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-95-f1cd5c4699da>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[1;33m[\u001b[0m\u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mvec\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mvec\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdocvecs\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32m<ipython-input-95-f1cd5c4699da>\u001b[0m in \u001b[0;36m<listcomp>\u001b[1;34m(.0)\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[1;33m[\u001b[0m\u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mvec\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mvec\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdocvecs\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32mc:\\users\\admin\\venv\\lib\\site-packages\\gensim\\models\\keyedvectors.py\u001b[0m in \u001b[0;36m__getitem__\u001b[1;34m(self, index)\u001b[0m\n\u001b[0;32m   1576\u001b[0m                 \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvectors_docs\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_int_index\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mindex\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdoctags\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmax_rawint\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1577\u001b[0m             \u001b[1;32mreturn\u001b[0m \u001b[0mvstack\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mindex\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1578\u001b[1;33m         \u001b[1;32mraise\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"tag '%s' not seen in training corpus/invalid\"\u001b[0m \u001b[1;33m%\u001b[0m \u001b[0mindex\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1579\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1580\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__contains__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyError\u001b[0m: \"tag '3' not seen in training corpus/invalid\""
     ]
    }
   ],
   "source": [
    "[print(vec) for vec in model.docvecs]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Topic Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. LDA - Latent Dirichlet Allocation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_20newsgroups\n",
    "\n",
    "cats = ['rec.sport.hockey', 'comp.graphics', 'sci.space']\n",
    "newsgroups_train = fetch_20newsgroups(subset='train', categories=cats)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((1777,), (1777,))"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "newsgroups_train.filenames.shape, newsgroups_train.target.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's look at sample data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "From: golchowy@alchemy.chem.utoronto.ca (Gerald Olchowy)\n",
      "Subject: Re: Flames Truly Brutal in Loss\n",
      "Organization: University of Toronto Chemistry Department\n",
      "Lines: 11\n",
      "\n",
      "In article <vzhivov.735193129@cunews> vzhivov@superior.carleton.ca (Vladimir Zhivov) writes:\n",
      ">As the subject suggests the Flames were not impressive this afternoon,\n",
      ">dropping a 6-3 decision to the LA Kings. Most of the Flames neglected\n",
      ">to show up, especially in their own zone, as the Kings hit at least\n",
      ">five posts! The Flames best line was probably\n",
      "\n",
      "Mike Vernon is now 3 wins 11 losses plus that All-Star game debacle in\n",
      "afternoon games during his career...with another afternoon game with\n",
      "Los Angeles next Sunday...has the ABC deal doomed the Flames?\n",
      "\n",
      "Gerald\n",
      "\n",
      "From: tdawson@engin.umich.edu (Chris Herringshaw)\n",
      "Subject: WingCommanderII Graphics\n",
      "Organization: University of Michigan Engineering, Ann Arbor\n",
      "Lines: 8\n",
      "Distribution: world\n",
      "NNTP-Posting-Host: antithesis.engin.umich.edu\n",
      "\n",
      "   I was wondering if anyone knows where I can get more information about\n",
      "the graphics in the WingCommander series, and the RealSpace system they use.\n",
      "I think it's really awesome, and wouldn't mind being able to use similar\n",
      "features in programs.  Thanks in advance.\n",
      "\n",
      "\n",
      "Daemon\n",
      "\n",
      "\n",
      "From: kjenks@gothamcity.jsc.nasa.gov\n",
      "Subject: Space Station Redesign, JSC Alternative #4\n",
      "Organization: NASA/JSC/GM2, Space Shuttle Program Office \n",
      "X-Newsreader: TIN [version 1.1 PL8]\n",
      "Lines: 71\n",
      "\n",
      "I have 19 (2 MB worth!) uuencode'd GIF images contain charts outlining\n",
      "one of the many alternative Space Station designs being considered in\n",
      "Crystal City.  Mr. Mark Holderman works down the hall from me, and can\n",
      "be reached for comment at (713) 483-1317, or via e-mail at\n",
      "mholderm@jscprofs.nasa.gov.\n",
      "\n",
      "Mark proposed this design, which he calls \"Geode\" (\"rough on the\n",
      "outside, but a gem on the inside\") or the \"ET Strongback with\n",
      "integrated hab modules and centrifuge.\"  As you can see from file\n",
      "geodeA.gif, it uses a Space Shuttle External Tank (ET) in place of much\n",
      "of the truss which is currently part of Space Station Freedom.  The\n",
      "white track on the outside of the ET is used by the Station Remonte\n",
      "Manipulator System (SRMS) and by the Reaction Control System (RCS)\n",
      "pod.  This allows the RCS pod to move along the track so that thrusting\n",
      "can occur near the center of gravity (CG) of the Station as the mass\n",
      "properties of the Station change during assembly.\n",
      "\n",
      "The inline module design allows the Shuttle to dock more easily because\n",
      "it can approach closer to the Station's CG and at a structurally strong\n",
      "part of the Station.  In the current SSF design, docking forces are\n",
      "limited to 400 pounds, which seriously constrains the design of the\n",
      "docking system.\n",
      "\n",
      "The ET would have a hatch installed pre-flight, with little additional\n",
      "launch mass.  We've always had the ability to put an ET into orbit\n",
      "(contrary to some rumors which have circulated here), but we've never\n",
      "had a reason to do it, while we have had some good reasons not to\n",
      "(performance penalties, control, debris generation, and eventual\n",
      "de-orbit and impact footprint).  Once on-orbit, we would vent the\n",
      "residual H2.  The ET insulation (SOFI) either a) erodes on-orbit from\n",
      "impact with atomic Oxygen, or b) stays where it is, and we deploy a\n",
      "Kevlar sheath around it to protect it and keep it from contaminating\n",
      "the local space environment.  Option b) has the advantage of providing\n",
      "further micrometeor protection.  The ET is incredibly strong (remember,\n",
      "it supports the whole stack during launch), and could serve as the\n",
      "nucleus for a much more ambitious design as budget permits.\n",
      "\n",
      "The white module at the end of ET contains a set of Control Moment\n",
      "Gyros to be used for attitude control, while the RCS will be used\n",
      "for gyro desaturation.  The module also contains a de-orbit system\n",
      "which can be used at the end of the Station's life to perform a\n",
      "controlled de-orbit (so we don't kill any more kangaroos, like we\n",
      "did with Skylab).\n",
      "\n",
      "The centrifuge, which has the same volume as a hab module, could be\n",
      "used for long-term studies of the effects of lunar or martian gravity\n",
      "on humans.  The centrifuge will be used as a momentum storage device\n",
      "for the whole attitude control system.  The centrifuge is mounted on\n",
      "one of the modules, opposite the ET and the solar panels.\n",
      "\n",
      "This design uses most of the existing SSF designs for electrical,\n",
      "data and communication systems, getting leverage from the SSF work\n",
      "done to date.\n",
      "\n",
      "Mark proposed this design at Joe Shea's committee in Crystal City,\n",
      "and he reports that he was warmly received.  However, the rumors\n",
      "I hear say that a design based on a wingless Space Shuttle Orbiter\n",
      "seems more likely.\n",
      "\n",
      "Please note that this text is my interpretation of Mark's design;\n",
      "you should see his notes in the GIF files.  \n",
      "\n",
      "Instead of posting a 2 MB file to sci.space, I tried to post these for\n",
      "anon-FTP in ames.arc.nasa.gov, but it was out of storage space.  I'll\n",
      "let you all know when I get that done.\n",
      "\n",
      "-- Ken Jenks, NASA/JSC/GM2, Space Shuttle Program Office\n",
      "      kjenks@gothamcity.jsc.nasa.gov  (713) 483-4368\n",
      "\n",
      "     \"...Development of the space station is as inevitable as \n",
      "      the rising of the sun.\" -- Wernher von Braun\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(\"\\n\".join(newsgroups_train.data[:3]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['rec.sport.hockey', 'comp.graphics', 'sci.space'], dtype='<U16')"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "np.array(newsgroups_train.target_names)[newsgroups_train.target[:3]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Text Vectorization\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer\n",
    "\n",
    "no_features = 1000\n",
    "documents = newsgroups_train.data\n",
    "\n",
    "# NMF is able to use tf-idf\n",
    "#tfidf_vectorizer = TfidfVectorizer(max_df=0.95, min_df=2, max_features=no_features, stop_words='english')\n",
    "#tfidf = tfidf_vectorizer.fit_transform(documents)\n",
    "#tfidf_feature_names = tfidf_vectorizer.get_feature_names()\n",
    "\n",
    "# LDA can only use raw term counts for LDA because it is a probabilistic graphical model\n",
    "tf_vectorizer = CountVectorizer(max_df=0.95, min_df=2, max_features=no_features, stop_words='english')\n",
    "tf = tf_vectorizer.fit_transform(documents)\n",
    "tf_feature_names = tf_vectorizer.get_feature_names()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 0 0 ... 1 0 0]\n",
      " [0 0 0 ... 0 0 0]\n",
      " [0 0 0 ... 0 0 0]\n",
      " ...\n",
      " [0 1 0 ... 0 0 0]\n",
      " [0 0 0 ... 0 0 0]\n",
      " [2 0 3 ... 0 0 0]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(1777, 1000)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(tf.toarray())\n",
    "tf.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1000\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['00',\n",
       " '000',\n",
       " '01',\n",
       " '02',\n",
       " '03',\n",
       " '04',\n",
       " '10',\n",
       " '100',\n",
       " '11',\n",
       " '12',\n",
       " '128',\n",
       " '13',\n",
       " '14',\n",
       " '15',\n",
       " '16',\n",
       " '17',\n",
       " '18',\n",
       " '19',\n",
       " '1992',\n",
       " '1993',\n",
       " '1993apr20',\n",
       " '1st',\n",
       " '20',\n",
       " '200',\n",
       " '21',\n",
       " '22',\n",
       " '23',\n",
       " '24',\n",
       " '25',\n",
       " '256',\n",
       " '26',\n",
       " '27',\n",
       " '28',\n",
       " '29',\n",
       " '2nd',\n",
       " '30',\n",
       " '300',\n",
       " '31',\n",
       " '32',\n",
       " '33',\n",
       " '34',\n",
       " '35',\n",
       " '36',\n",
       " '37',\n",
       " '38',\n",
       " '39',\n",
       " '3d',\n",
       " '3rd',\n",
       " '40',\n",
       " '41',\n",
       " '42',\n",
       " '43',\n",
       " '44',\n",
       " '45',\n",
       " '46',\n",
       " '47',\n",
       " '48',\n",
       " '49',\n",
       " '50',\n",
       " '500',\n",
       " '51',\n",
       " '52',\n",
       " '53',\n",
       " '54',\n",
       " '55',\n",
       " '56',\n",
       " '57',\n",
       " '58',\n",
       " '59',\n",
       " '60',\n",
       " '61',\n",
       " '62',\n",
       " '63',\n",
       " '64',\n",
       " '65',\n",
       " '66',\n",
       " '67',\n",
       " '68',\n",
       " '69',\n",
       " '70',\n",
       " '71',\n",
       " '72',\n",
       " '73',\n",
       " '74',\n",
       " '75',\n",
       " '76',\n",
       " '77',\n",
       " '78',\n",
       " '79',\n",
       " '80',\n",
       " '81',\n",
       " '82',\n",
       " '83',\n",
       " '84',\n",
       " '85',\n",
       " '86',\n",
       " '87',\n",
       " '90',\n",
       " '91',\n",
       " '92',\n",
       " '93',\n",
       " '95',\n",
       " '__',\n",
       " '___',\n",
       " '____',\n",
       " '_____',\n",
       " 'abc',\n",
       " 'able',\n",
       " 'ac',\n",
       " 'acad3',\n",
       " 'access',\n",
       " 'according',\n",
       " 'actually',\n",
       " 'adams',\n",
       " 'add',\n",
       " 'added',\n",
       " 'address',\n",
       " 'advance',\n",
       " 'ago',\n",
       " 'agree',\n",
       " 'ahl',\n",
       " 'air',\n",
       " 'al',\n",
       " 'alaska',\n",
       " 'alchemy',\n",
       " 'algorithm',\n",
       " 'allen',\n",
       " 'allow',\n",
       " 'american',\n",
       " 'ames',\n",
       " 'amiga',\n",
       " 'analysis',\n",
       " 'andrew',\n",
       " 'angeles',\n",
       " 'animation',\n",
       " 'announced',\n",
       " 'anonymous',\n",
       " 'answer',\n",
       " 'answers',\n",
       " 'anybody',\n",
       " 'apple',\n",
       " 'applications',\n",
       " 'appreciated',\n",
       " 'apr',\n",
       " 'april',\n",
       " 'archive',\n",
       " 'area',\n",
       " 'aren',\n",
       " 'arena',\n",
       " 'article',\n",
       " 'ask',\n",
       " 'asked',\n",
       " 'astro',\n",
       " 'astronomy',\n",
       " 'atmosphere',\n",
       " 'att',\n",
       " 'au',\n",
       " 'aurora',\n",
       " 'australia',\n",
       " 'author',\n",
       " 'available',\n",
       " 'away',\n",
       " 'baalke',\n",
       " 'bad',\n",
       " 'base',\n",
       " 'baseball',\n",
       " 'based',\n",
       " 'basically',\n",
       " 'bay',\n",
       " 'bbs',\n",
       " 'beat',\n",
       " 'believe',\n",
       " 'berkeley',\n",
       " 'best',\n",
       " 'better',\n",
       " 'big',\n",
       " 'biggest',\n",
       " 'billion',\n",
       " 'bit',\n",
       " 'bitnet',\n",
       " 'black',\n",
       " 'blue',\n",
       " 'blues',\n",
       " 'board',\n",
       " 'bob',\n",
       " 'body',\n",
       " 'book',\n",
       " 'books',\n",
       " 'bos',\n",
       " 'boston',\n",
       " 'box',\n",
       " 'brian',\n",
       " 'brown',\n",
       " 'bruins',\n",
       " 'buf',\n",
       " 'buffalo',\n",
       " 'build',\n",
       " 'building',\n",
       " 'built',\n",
       " 'business',\n",
       " 'buy',\n",
       " 'ca',\n",
       " 'cal',\n",
       " 'calgary',\n",
       " 'california',\n",
       " 'called',\n",
       " 'came',\n",
       " 'canada',\n",
       " 'canadian',\n",
       " 'canucks',\n",
       " 'caps',\n",
       " 'captain',\n",
       " 'card',\n",
       " 'carry',\n",
       " 'case',\n",
       " 'cc',\n",
       " 'cd',\n",
       " 'centaur',\n",
       " 'center',\n",
       " 'certainly',\n",
       " 'ch',\n",
       " 'chance',\n",
       " 'change',\n",
       " 'check',\n",
       " 'chem',\n",
       " 'chi',\n",
       " 'chicago',\n",
       " 'chris',\n",
       " 'circle',\n",
       " 'city',\n",
       " 'clark',\n",
       " 'clarkson',\n",
       " 'close',\n",
       " 'club',\n",
       " 'cmu',\n",
       " 'coach',\n",
       " 'code',\n",
       " 'college',\n",
       " 'color',\n",
       " 'colorado',\n",
       " 'colors',\n",
       " 'columbia',\n",
       " 'com',\n",
       " 'come',\n",
       " 'comes',\n",
       " 'comet',\n",
       " 'command',\n",
       " 'commercial',\n",
       " 'communications',\n",
       " 'comp',\n",
       " 'company',\n",
       " 'complete',\n",
       " 'computer',\n",
       " 'computing',\n",
       " 'conference',\n",
       " 'contact',\n",
       " 'control',\n",
       " 'convert',\n",
       " 'copy',\n",
       " 'corp',\n",
       " 'cost',\n",
       " 'costs',\n",
       " 'couldn',\n",
       " 'couple',\n",
       " 'course',\n",
       " 'coverage',\n",
       " 'craig',\n",
       " 'cs',\n",
       " 'cso',\n",
       " 'cunixb',\n",
       " 'cup',\n",
       " 'current',\n",
       " 'currently',\n",
       " 'cview',\n",
       " 'dare',\n",
       " 'data',\n",
       " 'date',\n",
       " 'dave',\n",
       " 'david',\n",
       " 'day',\n",
       " 'days',\n",
       " 'dc',\n",
       " 'deal',\n",
       " 'dec',\n",
       " 'defense',\n",
       " 'defenseman',\n",
       " 'defensive',\n",
       " 'den',\n",
       " 'department',\n",
       " 'dept',\n",
       " 'design',\n",
       " 'designed',\n",
       " 'det',\n",
       " 'details',\n",
       " 'detroit',\n",
       " 'developed',\n",
       " 'development',\n",
       " 'devils',\n",
       " 'did',\n",
       " 'didn',\n",
       " 'different',\n",
       " 'digex',\n",
       " 'directory',\n",
       " 'disk',\n",
       " 'display',\n",
       " 'distribution',\n",
       " 'division',\n",
       " 'does',\n",
       " 'doesn',\n",
       " 'doing',\n",
       " 'don',\n",
       " 'dos',\n",
       " 'double',\n",
       " 'doug',\n",
       " 'dr',\n",
       " 'draft',\n",
       " 'drawing',\n",
       " 'driver',\n",
       " 'dseg',\n",
       " 'early',\n",
       " 'earth',\n",
       " 'easily',\n",
       " 'easy',\n",
       " 'ed',\n",
       " 'edm',\n",
       " 'edmonton',\n",
       " 'edu',\n",
       " 'email',\n",
       " 'end',\n",
       " 'energy',\n",
       " 'eng',\n",
       " 'engineering',\n",
       " 'environment',\n",
       " 'eos',\n",
       " 'eric',\n",
       " 'ericsson',\n",
       " 'especially',\n",
       " 'espn',\n",
       " 'european',\n",
       " 'europeans',\n",
       " 'exactly',\n",
       " 'example',\n",
       " 'excellent',\n",
       " 'expansion',\n",
       " 'expected',\n",
       " 'exploration',\n",
       " 'face',\n",
       " 'facility',\n",
       " 'fact',\n",
       " 'fan',\n",
       " 'fans',\n",
       " 'faq',\n",
       " 'far',\n",
       " 'fast',\n",
       " 'fax',\n",
       " 'features',\n",
       " 'feel',\n",
       " 'fi',\n",
       " 'field',\n",
       " 'file',\n",
       " 'files',\n",
       " 'final',\n",
       " 'finally',\n",
       " 'finals',\n",
       " 'finland',\n",
       " 'flames',\n",
       " 'flight',\n",
       " 'flyers',\n",
       " 'fnal',\n",
       " 'following',\n",
       " 'force',\n",
       " 'form',\n",
       " 'format',\n",
       " 'formats',\n",
       " 'francis',\n",
       " 'fred',\n",
       " 'free',\n",
       " 'ftp',\n",
       " 'fuel',\n",
       " 'funding',\n",
       " 'future',\n",
       " 'gainey',\n",
       " 'galileo',\n",
       " 'game',\n",
       " 'games',\n",
       " 'gary',\n",
       " 'gatech',\n",
       " 'general',\n",
       " 'george',\n",
       " 'gerald',\n",
       " 'germany',\n",
       " 'gets',\n",
       " 'getting',\n",
       " 'gif',\n",
       " 'gilmour',\n",
       " 'given',\n",
       " 'gld',\n",
       " 'gm',\n",
       " 'gmt',\n",
       " 'goal',\n",
       " 'goalie',\n",
       " 'goals',\n",
       " 'goes',\n",
       " 'going',\n",
       " 'good',\n",
       " 'got',\n",
       " 'gov',\n",
       " 'government',\n",
       " 'grant',\n",
       " 'graphics',\n",
       " 'gravity',\n",
       " 'great',\n",
       " 'green',\n",
       " 'greg',\n",
       " 'gretzky',\n",
       " 'ground',\n",
       " 'group',\n",
       " 'groups',\n",
       " 'guess',\n",
       " 'guy',\n",
       " 'habs',\n",
       " 'half',\n",
       " 'hall',\n",
       " 'hand',\n",
       " 'har',\n",
       " 'hard',\n",
       " 'hardware',\n",
       " 'hartford',\n",
       " 'haven',\n",
       " 'having',\n",
       " 'hawks',\n",
       " 'head',\n",
       " 'hear',\n",
       " 'heard',\n",
       " 'help',\n",
       " 'henry',\n",
       " 'hi',\n",
       " 'higgins',\n",
       " 'high',\n",
       " 'history',\n",
       " 'hockey',\n",
       " 'home',\n",
       " 'hope',\n",
       " 'host',\n",
       " 'hours',\n",
       " 'hp',\n",
       " 'human',\n",
       " 'ibm',\n",
       " 'ice',\n",
       " 'idea',\n",
       " 'ideas',\n",
       " 'ii',\n",
       " 'image',\n",
       " 'images',\n",
       " 'imagine',\n",
       " 'imaging',\n",
       " 'important',\n",
       " 'include',\n",
       " 'included',\n",
       " 'includes',\n",
       " 'including',\n",
       " 'info',\n",
       " 'information',\n",
       " 'inside',\n",
       " 'instead',\n",
       " 'institute',\n",
       " 'interested',\n",
       " 'interesting',\n",
       " 'interface',\n",
       " 'international',\n",
       " 'internet',\n",
       " 'involved',\n",
       " 'islanders',\n",
       " 'isn',\n",
       " 'issue',\n",
       " 'ists',\n",
       " 'jagr',\n",
       " 'james',\n",
       " 'jersey',\n",
       " 'jet',\n",
       " 'jets',\n",
       " 'jim',\n",
       " 'job',\n",
       " 'joe',\n",
       " 'john',\n",
       " 'jose',\n",
       " 'joseph',\n",
       " 'jpeg',\n",
       " 'jpl',\n",
       " 'jsc',\n",
       " 'jupiter',\n",
       " 'just',\n",
       " 'keith',\n",
       " 'kelvin',\n",
       " 'ken',\n",
       " 'kevin',\n",
       " 'key',\n",
       " 'keywords',\n",
       " 'kind',\n",
       " 'king',\n",
       " 'kings',\n",
       " 'km',\n",
       " 'know',\n",
       " 'known',\n",
       " 'la',\n",
       " 'lab',\n",
       " 'laboratory',\n",
       " 'language',\n",
       " 'large',\n",
       " 'late',\n",
       " 'later',\n",
       " 'launch',\n",
       " 'launched',\n",
       " 'laurentian',\n",
       " 'lead',\n",
       " 'leafs',\n",
       " 'league',\n",
       " 'left',\n",
       " 'lemieux',\n",
       " 'let',\n",
       " 'level',\n",
       " 'library',\n",
       " 'life',\n",
       " 'light',\n",
       " 'like',\n",
       " 'likely',\n",
       " 'lindros',\n",
       " 'line',\n",
       " 'list',\n",
       " 'little',\n",
       " 'live',\n",
       " 'll',\n",
       " 'local',\n",
       " 'long',\n",
       " 'look',\n",
       " 'looked',\n",
       " 'looking',\n",
       " 'looks',\n",
       " 'los',\n",
       " 'loss',\n",
       " 'lost',\n",
       " 'lot',\n",
       " 'louis',\n",
       " 'low',\n",
       " 'lunar',\n",
       " 'mac',\n",
       " 'machine',\n",
       " 'mail',\n",
       " 'major',\n",
       " 'make',\n",
       " 'makes',\n",
       " 'making',\n",
       " 'man',\n",
       " 'management',\n",
       " 'maple',\n",
       " 'mark',\n",
       " 'market',\n",
       " 'mars',\n",
       " 'mass',\n",
       " 'material',\n",
       " 'math',\n",
       " 'matter',\n",
       " 'maybe',\n",
       " 'maynard',\n",
       " 'mccall',\n",
       " 'mean',\n",
       " 'means',\n",
       " 'media',\n",
       " 'memory',\n",
       " 'mention',\n",
       " 'message',\n",
       " 'michael',\n",
       " 'mike',\n",
       " 'mil',\n",
       " 'military',\n",
       " 'million',\n",
       " 'min',\n",
       " 'mind',\n",
       " 'minnesota',\n",
       " 'minor',\n",
       " 'minutes',\n",
       " 'mission',\n",
       " 'missions',\n",
       " 'mit',\n",
       " 'mode',\n",
       " 'model',\n",
       " 'models',\n",
       " 'mon',\n",
       " 'money',\n",
       " 'months',\n",
       " 'montreal',\n",
       " 'moon',\n",
       " 'ms',\n",
       " 'murray',\n",
       " 'na',\n",
       " 'names',\n",
       " 'nasa',\n",
       " 'national',\n",
       " 'navy',\n",
       " 'ncsu',\n",
       " 'near',\n",
       " 'need',\n",
       " 'needed',\n",
       " 'needs',\n",
       " 'net',\n",
       " 'network',\n",
       " 'new',\n",
       " 'news',\n",
       " 'newsgroup',\n",
       " 'newsreader',\n",
       " 'nhl',\n",
       " 'nice',\n",
       " 'nick',\n",
       " 'night',\n",
       " 'nj',\n",
       " 'nl',\n",
       " 'nntp',\n",
       " 'non',\n",
       " 'norris',\n",
       " 'north',\n",
       " 'note',\n",
       " 'nsmca',\n",
       " 'number',\n",
       " 'numbers',\n",
       " 'ny',\n",
       " 'nyi',\n",
       " 'nyr',\n",
       " 'object',\n",
       " 'objects',\n",
       " 'observatory',\n",
       " 'office',\n",
       " 'oh',\n",
       " 'oilers',\n",
       " 'old',\n",
       " 'online',\n",
       " 'ontario',\n",
       " 'open',\n",
       " 'operations',\n",
       " 'opinion',\n",
       " 'opinions',\n",
       " 'orbit',\n",
       " 'orbital',\n",
       " 'orbiter',\n",
       " 'order',\n",
       " 'org',\n",
       " 'original',\n",
       " 'ott',\n",
       " 'ottawa',\n",
       " 'oz',\n",
       " 'p2',\n",
       " 'package',\n",
       " 'paper',\n",
       " 'pass',\n",
       " 'past',\n",
       " 'pat',\n",
       " 'patrick',\n",
       " 'paul',\n",
       " 'pay',\n",
       " 'pc',\n",
       " 'penalties',\n",
       " 'penalty',\n",
       " 'penguins',\n",
       " 'pens',\n",
       " 'people',\n",
       " 'period',\n",
       " 'person',\n",
       " 'peter',\n",
       " 'phi',\n",
       " 'philadelphia',\n",
       " 'phone',\n",
       " 'physics',\n",
       " 'pick',\n",
       " 'picture',\n",
       " 'pit',\n",
       " 'pitt',\n",
       " 'pittsburgh',\n",
       " 'place',\n",
       " 'plane',\n",
       " 'planet',\n",
       " 'planetary',\n",
       " 'play',\n",
       " 'played',\n",
       " 'player',\n",
       " 'players',\n",
       " 'playing',\n",
       " 'playoff',\n",
       " 'playoffs',\n",
       " 'plus',\n",
       " 'point',\n",
       " 'points',\n",
       " 'polygon',\n",
       " 'pool',\n",
       " 'position',\n",
       " 'possible',\n",
       " 'post',\n",
       " 'posted',\n",
       " 'posting',\n",
       " 'posts',\n",
       " 'postscript',\n",
       " 'pov',\n",
       " 'power',\n",
       " 'pp',\n",
       " 'prb',\n",
       " 'president',\n",
       " 'press',\n",
       " 'pretty',\n",
       " 'probably',\n",
       " 'probe',\n",
       " 'problem',\n",
       " 'problems',\n",
       " 'process',\n",
       " 'processing',\n",
       " 'product',\n",
       " 'program',\n",
       " 'programming',\n",
       " 'programs',\n",
       " 'project',\n",
       " 'propulsion',\n",
       " 'provide',\n",
       " 'pt',\n",
       " 'pts',\n",
       " 'pub',\n",
       " 'public',\n",
       " 'puck',\n",
       " 'purdue',\n",
       " 'purpose',\n",
       " 'quality',\n",
       " 'que',\n",
       " 'quebec',\n",
       " 'question',\n",
       " 'questions',\n",
       " 'quicktime',\n",
       " 'quite',\n",
       " 'radio',\n",
       " 'radius',\n",
       " 'ramsey',\n",
       " 'rangers',\n",
       " 'rate',\n",
       " 'ray',\n",
       " 'read',\n",
       " 'reading',\n",
       " 'real',\n",
       " 'reality',\n",
       " 'really',\n",
       " 'reason',\n",
       " 'recchi',\n",
       " 'recently',\n",
       " 'record',\n",
       " 'red',\n",
       " 'reference',\n",
       " 'regular',\n",
       " 'related',\n",
       " 'remember',\n",
       " 'remote',\n",
       " 'reply',\n",
       " 'report',\n",
       " 'request',\n",
       " 'require',\n",
       " 'research',\n",
       " 'rest',\n",
       " 'result',\n",
       " 'results',\n",
       " 'return',\n",
       " 'richard',\n",
       " 'rick',\n",
       " 'right',\n",
       " 'robert',\n",
       " 'rochester',\n",
       " 'rocket',\n",
       " 'rockets',\n",
       " 'roger',\n",
       " 'ron',\n",
       " 'round',\n",
       " 'rpi',\n",
       " 'run',\n",
       " 'running',\n",
       " 'runs',\n",
       " 'russian',\n",
       " 'sabres',\n",
       " 'said',\n",
       " 'san',\n",
       " 'sas',\n",
       " 'satellite',\n",
       " 'satellites',\n",
       " 'saturn',\n",
       " 'saw',\n",
       " 'say',\n",
       " 'saying',\n",
       " 'says',\n",
       " 'scale',\n",
       " 'school',\n",
       " 'sci',\n",
       " 'science',\n",
       " 'sciences',\n",
       " 'scientific',\n",
       " 'score',\n",
       " 'scored',\n",
       " 'scorer',\n",
       " 'scoring',\n",
       " 'scott',\n",
       " 'screen',\n",
       " 'se',\n",
       " 'season',\n",
       " 'second',\n",
       " 'section',\n",
       " 'seen',\n",
       " 'selanne',\n",
       " 'send',\n",
       " 'sent',\n",
       " 'series',\n",
       " 'server',\n",
       " 'service',\n",
       " 'services',\n",
       " 'set',\n",
       " 'sfu',\n",
       " 'sgi',\n",
       " 'shareware',\n",
       " 'sharks',\n",
       " 'short',\n",
       " 'shot',\n",
       " 'shots',\n",
       " 'shuttle',\n",
       " 'similar',\n",
       " 'simple',\n",
       " 'simply',\n",
       " 'single',\n",
       " 'site',\n",
       " 'sites',\n",
       " 'size',\n",
       " 'sj',\n",
       " 'sky',\n",
       " 'small',\n",
       " 'smith',\n",
       " 'soderstrom',\n",
       " 'software',\n",
       " 'solar',\n",
       " 'solid',\n",
       " 'sorry',\n",
       " 'sort',\n",
       " 'source',\n",
       " 'sources',\n",
       " 'space',\n",
       " 'spacecraft',\n",
       " 'special',\n",
       " 'specific',\n",
       " 'speed',\n",
       " 'spencer',\n",
       " 'sphere',\n",
       " 'sport',\n",
       " 'st',\n",
       " 'stage',\n",
       " 'standard',\n",
       " 'stanford',\n",
       " 'stanley',\n",
       " 'star',\n",
       " 'stars',\n",
       " 'start',\n",
       " 'started',\n",
       " 'state',\n",
       " 'station',\n",
       " 'stats',\n",
       " 'steve',\n",
       " 'stl',\n",
       " 'stop',\n",
       " 'stuff',\n",
       " 'summary',\n",
       " 'sun',\n",
       " 'support',\n",
       " 'sure',\n",
       " 'surface',\n",
       " 'sweden',\n",
       " 'systems',\n",
       " 'taking',\n",
       " 'talk',\n",
       " 'talking',\n",
       " 'tampa',\n",
       " 'tar',\n",
       " 'tb',\n",
       " 'team',\n",
       " 'teams',\n",
       " 'technical',\n",
       " 'technology',\n",
       " 'telescope',\n",
       " 'tell',\n",
       " 'test',\n",
       " 'texas',\n",
       " 'text',\n",
       " 'thanks',\n",
       " 'theory',\n",
       " 'thing',\n",
       " 'things',\n",
       " 'think',\n",
       " 'thomas',\n",
       " 'thought',\n",
       " 'ti',\n",
       " 'tie',\n",
       " 'tiff',\n",
       " 'time',\n",
       " 'times',\n",
       " 'tin',\n",
       " 'titan',\n",
       " 'today',\n",
       " 'tom',\n",
       " 'tommy',\n",
       " 'took',\n",
       " 'tools',\n",
       " 'tor',\n",
       " 'toronto',\n",
       " 'total',\n",
       " 'trade',\n",
       " 'traded',\n",
       " 'tried',\n",
       " 'true',\n",
       " 'try',\n",
       " 'trying',\n",
       " 'tv',\n",
       " 'type',\n",
       " 'uiuc',\n",
       " 'uk',\n",
       " 'umd',\n",
       " 'umich',\n",
       " 'unc',\n",
       " 'understand',\n",
       " 'univ',\n",
       " 'universe',\n",
       " 'university',\n",
       " 'unix',\n",
       " 'unless',\n",
       " 'upenn',\n",
       " 'usa',\n",
       " 'use',\n",
       " 'used',\n",
       " 'useful',\n",
       " 'usenet',\n",
       " 'user',\n",
       " 'users',\n",
       " 'uses',\n",
       " 'using',\n",
       " 'utoronto',\n",
       " 'uucp',\n",
       " 'uvic',\n",
       " 'uwo',\n",
       " 'value',\n",
       " 'van',\n",
       " 'vancouver',\n",
       " 'various',\n",
       " 'vax',\n",
       " 've',\n",
       " 'vehicle',\n",
       " 'venus',\n",
       " 'version',\n",
       " 'vesa',\n",
       " 'vga',\n",
       " 'video',\n",
       " 'view',\n",
       " 'viewer',\n",
       " 'virginia',\n",
       " 'virtual',\n",
       " 'vms',\n",
       " 'volume',\n",
       " 'vs',\n",
       " 'want',\n",
       " 'wanted',\n",
       " 'washington',\n",
       " 'wasn',\n",
       " 'watch',\n",
       " 'way',\n",
       " 'wayne',\n",
       " 'week',\n",
       " 'went',\n",
       " 'western',\n",
       " 'white',\n",
       " 'win',\n",
       " 'windows',\n",
       " 'wings',\n",
       " 'winner',\n",
       " 'winning',\n",
       " 'winnipeg',\n",
       " 'wins',\n",
       " 'won',\n",
       " 'work',\n",
       " 'working',\n",
       " 'works',\n",
       " 'world',\n",
       " 'worth',\n",
       " 'wouldn',\n",
       " 'write',\n",
       " 'writes',\n",
       " 'written',\n",
       " 'wrong',\n",
       " 'wrote',\n",
       " 'xv',\n",
       " 'year',\n",
       " 'years',\n",
       " 'yes',\n",
       " 'york',\n",
       " 'young',\n",
       " 'zone',\n",
       " 'zoo',\n",
       " 'zoology']"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(len(tf_feature_names))\n",
    "tf_feature_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\admin\\venv\\lib\\site-packages\\sklearn\\decomposition\\online_lda.py:314: DeprecationWarning: n_topics has been renamed to n_components in version 0.19 and will be removed in 0.21\n",
      "  DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.decomposition import LatentDirichletAllocation\n",
    "\n",
    "no_topics = 3\n",
    "\n",
    "# Run NMF\n",
    "#nmf = NMF(n_components=no_topics, random_state=1, alpha=.1, l1_ratio=.5, init='nndsvd').fit(tfidf)\n",
    "\n",
    "# Run LDA\n",
    "lda = LatentDirichletAllocation(n_topics=no_topics, max_iter=5, learning_method='online', learning_offset=50.,random_state=0).fit(tf)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Topic 0:\n",
      "space edu nasa image graphics\n",
      "Topic 1:\n",
      "edu writes article com posting\n",
      "Topic 2:\n",
      "team hockey game play 10\n"
     ]
    }
   ],
   "source": [
    "def display_topics(model, feature_names, no_top_words):\n",
    "    for topic_idx, topic in enumerate(model.components_):\n",
    "        print(\"Topic %d:\" % (topic_idx))\n",
    "        print(\" \".join([feature_names[i]\n",
    "                        for i in topic.argsort()[:-no_top_words - 1:-1]]))\n",
    "\n",
    "no_top_words = 5\n",
    "\n",
    "#display_topics(nmf, tfidf_feature_names, no_top_words)\n",
    "display_topics(lda, tf_feature_names, no_top_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.Non-negative matrix factorization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### SVD "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We would clearly expect that the words that appear most frequently in one topic would appear less frequently in the other - otherwise that word wouldn't make a good choice to separate out the two topics. Therefore, we expect the topics to be **orthogonal**.\n",
    "\n",
    "The SVD algorithm factorizes a matrix into one matrix with **orthogonal columns** and one with **orthogonal rows** (along with a diagonal matrix, which contains the **relative importance** of each factor).\n",
    "\n",
    "<img src=\"images/svd_fb.png\" alt=\"\" style=\"width: 80%\"/>\n",
    "(source: [Facebook Research: Fast Randomized SVD](https://research.fb.com/fast-randomized-svd/))\n",
    "\n",
    "SVD is an **exact decomposition**, since the matrices it creates are big enough to fully cover the original matrix. SVD is extremely widely used in ML, including:\n",
    "\n",
    "- collaborative filtering/recommendations ([winning entry for Netflix Prize](https://datajobs.com/data-science-repo/Recommender-Systems-%5BNetflix%5D.pdf))\n",
    "- calculate Moore-Penrose pseudoinverse\n",
    "- data compression\n",
    "- principal component analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"images/nmf_doc.png\" alt=\"NMF on documents\" style=\"width: 80%\"/>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"images/nmf_example.png\" alt=\"NMF on documents\" style=\"width: 80%\"/>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Text Vectorization\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer\n",
    "\n",
    "no_features = 1000\n",
    "documents = newsgroups_train.data\n",
    "\n",
    "#NMF is able to use tf-idf\n",
    "tfidf_vectorizer = TfidfVectorizer(max_df=0.95, min_df=2, max_features=no_features, stop_words='english')\n",
    "tfidf = tfidf_vectorizer.fit_transform(documents)\n",
    "tfidf_feature_names = tfidf_vectorizer.get_feature_names()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.decomposition import NMF\n",
    "\n",
    "no_topics = 5\n",
    "\n",
    "#Run NMF\n",
    "nmf_model = NMF(n_components=no_topics, random_state=1, alpha=.1, l1_ratio=.5, init='nndsvd')\n",
    "nmf = nmf_model.fit(tfidf)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "H1 = nmf_model.components_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'show_topics' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-18-1cf779cd9c0d>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mshow_topics\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mH1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'show_topics' is not defined"
     ]
    }
   ],
   "source": [
    "show_topics(H1)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}