code/ch08/ch08.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright (c) 2015, 2016 [Sebastian Raschka](sebastianraschka.com)\n",
    "\n",
    "https://github.com/rasbt/python-machine-learning-book\n",
    "\n",
    "[MIT License](https://github.com/rasbt/python-machine-learning-book/blob/master/LICENSE.txt)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Python Machine Learning - Code Examples"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chapter 8 - Applying Machine Learning To Sentiment Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that the optional watermark extension is a small IPython notebook plugin that I developed to make the code reproducible. You can just skip the following line(s)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sebastian Raschka \n",
      "last updated: 2016-09-29 \n",
      "\n",
      "CPython 3.5.2\n",
      "IPython 5.1.0\n",
      "\n",
      "numpy 1.11.1\n",
      "pandas 0.18.1\n",
      "matplotlib 1.5.1\n",
      "sklearn 0.18\n",
      "nltk 3.2.1\n"
     ]
    }
   ],
   "source": [
    "%load_ext watermark\n",
    "%watermark -a 'Sebastian Raschka' -u -d -v -p numpy,pandas,matplotlib,sklearn,nltk"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "*The use of `watermark` is optional. You can install this IPython extension via \"`pip install watermark`\". For more information, please see: https://github.com/rasbt/watermark.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Overview"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- [Obtaining the IMDb movie review dataset](#Obtaining-the-IMDb-movie-review-dataset)\n",
    "- [Introducing the bag-of-words model](#Introducing-the-bag-of-words-model)\n",
    "  - [Transforming words into feature vectors](#Transforming-words-into-feature-vectors)\n",
    "  - [Assessing word relevancy via term frequency-inverse document frequency](#Assessing-word-relevancy-via-term-frequency-inverse-document-frequency)\n",
    "  - [Cleaning text data](#Cleaning-text-data)\n",
    "  - [Processing documents into tokens](#Processing-documents-into-tokens)\n",
    "- [Training a logistic regression model for document classification](#Training-a-logistic-regression-model-for-document-classification)\n",
    "- [Working with bigger data – online algorithms and out-of-core learning](#Working-with-bigger-data-–-online-algorithms-and-out-of-core-learning)\n",
    "- [Summary](#Summary)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Added version check for recent scikit-learn 0.18 checks\n",
    "from distutils.version import LooseVersion as Version\n",
    "from sklearn import __version__ as sklearn_version"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Obtaining the IMDb movie review dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The IMDB movie review set can be downloaded from [http://ai.stanford.edu/~amaas/data/sentiment/](http://ai.stanford.edu/~amaas/data/sentiment/).\n",
    "After downloading the dataset, decompress the files.\n",
    "\n",
    "A) If you are working with Linux or MacOS X, open a new terminal windowm `cd` into the download directory and execute \n",
    "\n",
    "`tar -zxf aclImdb_v1.tar.gz`\n",
    "\n",
    "B) If you are working with Windows, download an archiver such as [7Zip](http://www.7-zip.org) to extract the files from the download archive."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compatibility Note:\n",
    "\n",
    "I received an email from a reader who was having troubles with reading the movie review texts due to encoding issues. Typically, Python's default encoding is set to `'utf-8'`, which shouldn't cause troubles when running this IPython notebook. You can simply check the encoding on your machine by firing up a new Python interpreter from the command line terminal and execute\n",
    "\n",
    "    >>> import sys\n",
    "    >>> sys.getdefaultencoding()\n",
    "    \n",
    "If the returned result is **not** `'utf-8'`, you probably need to change your Python's encoding to `'utf-8'`, for example by typing `export PYTHONIOENCODING=utf8` in your terminal shell prior to running this IPython notebook. (Note that this is a temporary change, and it needs to be executed in the same shell that you'll use to launch `ipython notebook`.\n",
    "\n",
    "Alternatively, you can replace the lines \n",
    "\n",
    "    with open(os.path.join(path, file), 'r') as infile:\n",
    "    ...\n",
    "    pd.read_csv('./movie_data.csv')\n",
    "    ...\n",
    "    df.to_csv('./movie_data.csv', index=False)\n",
    "\n",
    "by \n",
    "\n",
    "    with open(os.path.join(path, file), 'r', encoding='utf-8') as infile:\n",
    "    ...\n",
    "    pd.read_csv('./movie_data.csv', encoding='utf-8')\n",
    "    ...\n",
    "    df.to_csv('./movie_data.csv', index=False, encoding='utf-8')\n",
    "    \n",
    "in the following cells to achieve the desired effect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "0%                          100%\n",
      "[##############################] | ETA: 00:00:00\n",
      "Total time elapsed: 00:09:04\n"
     ]
    }
   ],
   "source": [
    "import pyprind\n",
    "import pandas as pd\n",
    "import os\n",
    "\n",
    "# change the `basepath` to the directory of the\n",
    "# unzipped movie dataset\n",
    "\n",
    "#basepath = '/Users/Sebastian/Desktop/aclImdb/'\n",
    "basepath = './aclImdb'\n",
    "\n",
    "labels = {'pos': 1, 'neg': 0}\n",
    "pbar = pyprind.ProgBar(50000)\n",
    "df = pd.DataFrame()\n",
    "for s in ('test', 'train'):\n",
    "    for l in ('pos', 'neg'):\n",
    "        path = os.path.join(basepath, s, l)\n",
    "        for file in os.listdir(path):\n",
    "            with open(os.path.join(path, file), 'r', encoding='utf-8') as infile:\n",
    "                txt = infile.read()\n",
    "            df = df.append([[txt, labels[l]]], ignore_index=True)\n",
    "            pbar.update()\n",
    "df.columns = ['review', 'sentiment']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Shuffling the DataFrame:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "np.random.seed(0)\n",
    "df = df.reindex(np.random.permutation(df.index))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Optional: Saving the assembled data as CSV file:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df.to_csv('./movie_data.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>In 1974, the teenager Martha Moxley (Maggie Gr...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>OK... so... I really like Kris Kristofferson a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>***SPOILER*** Do not read this, if you think a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              review  sentiment\n",
       "0  In 1974, the teenager Martha Moxley (Maggie Gr...          1\n",
       "1  OK... so... I really like Kris Kristofferson a...          0\n",
       "2  ***SPOILER*** Do not read this, if you think a...          0"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df = pd.read_csv('./movie_data.csv')\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "### Note\n",
    "\n",
    "If you have problems with creating the `movie_data.csv` file in the previous chapter, you can find a download a zip archive at \n",
    "https://github.com/rasbt/python-machine-learning-book/tree/master/code/datasets/movie\n",
    "<hr>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Introducing the bag-of-words model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Transforming documents into feature vectors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By calling the fit_transform method on CountVectorizer, we just constructed the vocabulary of the bag-of-words model and transformed the following three sentences into sparse feature vectors:\n",
    "1. The sun is shining\n",
    "2. The weather is sweet\n",
    "3. The sun is shining, the weather is sweet, and one and one is two\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "count = CountVectorizer()\n",
    "docs = np.array([\n",
    "        'The sun is shining',\n",
    "        'The weather is sweet',\n",
    "        'The sun is shining, the weather is sweet, and one and one is two'])\n",
    "bag = count.fit_transform(docs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let us print the contents of the vocabulary to get a better understanding of the underlying concepts:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'one': 2, 'sweet': 5, 'the': 6, 'shining': 3, 'weather': 8, 'and': 0, 'two': 7, 'is': 1, 'sun': 4}\n"
     ]
    }
   ],
   "source": [
    "print(count.vocabulary_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see from executing the preceding command, the vocabulary is stored in a Python dictionary, which maps the unique words that are mapped to integer indices. Next let us print the feature vectors that we just created:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Each index position in the feature vectors shown here corresponds to the integer values that are stored as dictionary items in the CountVectorizer vocabulary. For example, the  rst feature at index position 0 resembles the count of the word and, which only occurs in the last document, and the word is at index position 1 (the 2nd feature in the document vectors) occurs in all three sentences. Those values in the feature vectors are also called the raw term frequencies: *tf (t,d)*—the number of times a term t occurs in a document *d*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 1 0 1 1 0 1 0 0]\n",
      " [0 1 0 0 0 1 1 0 1]\n",
      " [2 3 2 1 1 1 2 1 1]]\n"
     ]
    }
   ],
   "source": [
    "print(bag.toarray())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Assessing word relevancy via term frequency-inverse document frequency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.set_printoptions(precision=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When we are analyzing text data, we often encounter words that occur across multiple documents from both classes. Those frequently occurring words typically don't contain useful or discriminatory information. In this subsection, we will learn about a useful technique called term frequency-inverse document frequency (tf-idf) that can be used to downweight those frequently occurring words in the feature vectors. The tf-idf can be de ned as the product of the term frequency and the inverse document frequency:\n",
    "\n",
    "$$\\text{tf-idf}(t,d)=\\text{tf (t,d)}\\times \\text{idf}(t,d)$$\n",
    "\n",
    "Here the tf(t, d) is the term frequency that we introduced in the previous section,\n",
    "and the inverse document frequency *idf(t, d)* can be calculated as:\n",
    "\n",
    "$$\\text{idf}(t,d) = \\text{log}\\frac{n_d}{1+\\text{df}(d, t)},$$\n",
    "\n",
    "where $n_d$ is the total number of documents, and *df(d, t)* is the number of documents *d* that contain the term *t*. Note that adding the constant 1 to the denominator is optional and serves the purpose of assigning a non-zero value to terms that occur in all training samples; the log is used to ensure that low document frequencies are not given too much weight.\n",
    "\n",
    "Scikit-learn implements yet another transformer, the `TfidfTransformer`, that takes the raw term frequencies from `CountVectorizer` as input and transforms them into tf-idfs:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.    0.43  0.    0.56  0.56  0.    0.43  0.    0.  ]\n",
      " [ 0.    0.43  0.    0.    0.    0.56  0.43  0.    0.56]\n",
      " [ 0.5   0.45  0.5   0.19  0.19  0.19  0.3   0.25  0.19]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "\n",
    "tfidf = TfidfTransformer(use_idf=True, norm='l2', smooth_idf=True)\n",
    "print(tfidf.fit_transform(count.fit_transform(docs)).toarray())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we saw in the previous subsection, the word is had the largest term frequency in the 3rd document, being the most frequently occurring word. However, after transforming the same feature vector into tf-idfs, we see that the word is is\n",
    "now associated with a relatively small tf-idf (0.45) in document 3 since it is\n",
    "also contained in documents 1 and 2 and thus is unlikely to contain any useful, discriminatory information.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "However, if we'd manually calculated the tf-idfs of the individual terms in our feature vectors, we'd have noticed that the `TfidfTransformer` calculates the tf-idfs slightly differently compared to the standard textbook equations that we de ned earlier. The equations for the idf and tf-idf that were implemented in scikit-learn are:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$$\\text{idf} (t,d) = log\\frac{1 + n_d}{1 + \\text{df}(d, t)}$$\n",
    "\n",
    "The tf-idf equation that was implemented in scikit-learn is as follows:\n",
    "\n",
    "$$\\text{tf-idf}(t,d) = \\text{tf}(t,d) \\times (\\text{idf}(t,d)+1)$$\n",
    "\n",
    "While it is also more typical to normalize the raw term frequencies before calculating the tf-idfs, the `TfidfTransformer` normalizes the tf-idfs directly.\n",
    "\n",
    "By default (`norm='l2'`), scikit-learn's TfidfTransformer applies the L2-normalization, which returns a vector of length 1 by dividing an un-normalized feature vector *v* by its L2-norm:\n",
    "\n",
    "$$v_{\\text{norm}} = \\frac{v}{||v||_2} = \\frac{v}{\\sqrt{v_{1}^{2} + v_{2}^{2} + \\dots + v_{n}^{2}}} = \\frac{v}{\\big (\\sum_{i=1}^{n} v_{i}^{2}\\big)^\\frac{1}{2}}$$\n",
    "\n",
    "To make sure that we understand how TfidfTransformer works, let us walk\n",
    "through an example and calculate the tf-idf of the word is in the 3rd document.\n",
    "\n",
    "The word is has a term frequency of 3 (tf = 3) in document 3, and the document frequency of this term is 3 since the term is occurs in all three documents (df = 3). Thus, we can calculate the idf as follows:\n",
    "\n",
    "$$\\text{idf}(\"is\", d3) = log \\frac{1+3}{1+3} = 0$$\n",
    "\n",
    "Now in order to calculate the tf-idf, we simply need to add 1 to the inverse document frequency and multiply it by the term frequency:\n",
    "\n",
    "$$\\text{tf-idf}(\"is\",d3)= 3 \\times (0+1) = 3$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf-idf of term \"is\" = 3.00\n"
     ]
    }
   ],
   "source": [
    "tf_is = 3\n",
    "n_docs = 3\n",
    "idf_is = np.log((n_docs+1) / (3+1))\n",
    "tfidf_is = tf_is * (idf_is + 1)\n",
    "print('tf-idf of term \"is\" = %.2f' % tfidf_is)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we repeated these calculations for all terms in the 3rd document, we'd obtain the following tf-idf vectors: [3.39, 3.0, 3.39, 1.29, 1.29, 1.29, 2.0 , 1.69, 1.29]. However, we notice that the values in this feature vector are different from the values that we obtained from the TfidfTransformer that we used previously. The  nal step that we are missing in this tf-idf calculation is the L2-normalization, which can be applied as follows:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$$\\text{tfi-df}_{norm} = \\frac{[3.39, 3.0, 3.39, 1.29, 1.29, 1.29, 2.0 , 1.69, 1.29]}{\\sqrt{[3.39^2, 3.0^2, 3.39^2, 1.29^2, 1.29^2, 1.29^2, 2.0^2 , 1.69^2, 1.29^2]}}$$\n",
    "\n",
    "$$=[0.5, 0.45, 0.5, 0.19, 0.19, 0.19, 0.3, 0.25, 0.19]$$\n",
    "\n",
    "$$\\Rightarrow \\text{tfi-df}_{norm}(\"is\", d3) = 0.45$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, the results match the results returned by scikit-learn's `TfidfTransformer` (below). Since we now understand how tf-idfs are calculated, let us proceed to the next sections and apply those concepts to the movie review dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 3.39,  3.  ,  3.39,  1.29,  1.29,  1.29,  2.  ,  1.69,  1.29])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tfidf = TfidfTransformer(use_idf=True, norm=None, smooth_idf=True)\n",
    "raw_tfidf = tfidf.fit_transform(count.fit_transform(docs)).toarray()[-1]\n",
    "raw_tfidf "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.5 ,  0.45,  0.5 ,  0.19,  0.19,  0.19,  0.3 ,  0.25,  0.19])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l2_tfidf = raw_tfidf / np.sqrt(np.sum(raw_tfidf**2))\n",
    "l2_tfidf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cleaning text data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'is seven.<br /><br />Title (Brazil): Not Available'"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[0, 'review'][-50:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import re\n",
    "def preprocessor(text):\n",
    "    text = re.sub('<[^>]*>', '', text)\n",
    "    emoticons = re.findall('(?::|;|=)(?:-)?(?:\\)|\\(|D|P)', text)\n",
    "    text = re.sub('[\\W]+', ' ', text.lower()) +\\\n",
    "        ' '.join(emoticons).replace('-', '')\n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'is seven title brazil not available'"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preprocessor(df.loc[0, 'review'][-50:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'this is a test :) :( :)'"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preprocessor(\"</a>This :) is :( a test :-)!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df['review'] = df['review'].apply(preprocessor)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Processing documents into tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from nltk.stem.porter import PorterStemmer\n",
    "\n",
    "porter = PorterStemmer()\n",
    "\n",
    "def tokenizer(text):\n",
    "    return text.split()\n",
    "\n",
    "\n",
    "def tokenizer_porter(text):\n",
    "    return [porter.stem(word) for word in text.split()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['runners', 'like', 'running', 'and', 'thus', 'they', 'run']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tokenizer('runners like running and thus they run')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['runner', 'like', 'run', 'and', 'thu', 'they', 'run']"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tokenizer_porter('runners like running and thus they run')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /Users/Sebastian/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import nltk\n",
    "\n",
    "nltk.download('stopwords')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['runner', 'like', 'run', 'run', 'lot']"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from nltk.corpus import stopwords\n",
    "\n",
    "stop = stopwords.words('english')\n",
    "[w for w in tokenizer_porter('a runner likes running and runs a lot')[-10:]\n",
    "if w not in stop]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training a logistic regression model for document classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Strip HTML and punctuation to speed up the GridSearch later:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_train = df.loc[:25000, 'review'].values\n",
    "y_train = df.loc[:25000, 'sentiment'].values\n",
    "X_test = df.loc[25000:, 'review'].values\n",
    "y_test = df.loc[25000:, 'sentiment'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "if Version(sklearn_version) < '0.18':\n",
    "    from sklearn.grid_search import GridSearchCV\n",
    "else:\n",
    "    from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "tfidf = TfidfVectorizer(strip_accents=None,\n",
    "                        lowercase=False,\n",
    "                        preprocessor=None)\n",
    "\n",
    "param_grid = [{'vect__ngram_range': [(1, 1)],\n",
    "               'vect__stop_words': [stop, None],\n",
    "               'vect__tokenizer': [tokenizer, tokenizer_porter],\n",
    "               'clf__penalty': ['l1', 'l2'],\n",
    "               'clf__C': [1.0, 10.0, 100.0]},\n",
    "              {'vect__ngram_range': [(1, 1)],\n",
    "               'vect__stop_words': [stop, None],\n",
    "               'vect__tokenizer': [tokenizer, tokenizer_porter],\n",
    "               'vect__use_idf':[False],\n",
    "               'vect__norm':[None],\n",
    "               'clf__penalty': ['l1', 'l2'],\n",
    "               'clf__C': [1.0, 10.0, 100.0]},\n",
    "              ]\n",
    "\n",
    "lr_tfidf = Pipeline([('vect', tfidf),\n",
    "                     ('clf', LogisticRegression(random_state=0))])\n",
    "\n",
    "gs_lr_tfidf = GridSearchCV(lr_tfidf, param_grid,\n",
    "                           scoring='accuracy',\n",
    "                           cv=5,\n",
    "                           verbose=1,\n",
    "                           n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 48 candidates, totalling 240 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed: 43.9min\n",
      "[Parallel(n_jobs=-1)]: Done 192 tasks      | elapsed: 228.2min\n",
      "[Parallel(n_jobs=-1)]: Done 240 out of 240 | elapsed: 265.3min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=Pipeline(steps=[('vect', TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
       "        lowercase=False, max_df=1.0, max_features=None, min_df=1,\n",
       "        ngram_range=(1, 1), norm='l2', preprocessor=None, smooth_idf=True,\n",
       " ...nalty='l2', random_state=0, solver='liblinear', tol=0.0001,\n",
       "          verbose=0, warm_start=False))]),\n",
       "       fit_params={}, iid=True, n_jobs=-1,\n",
       "       param_grid=[{'vect__tokenizer': [<function tokenizer at 0x11851c6a8>, <function tokenizer_porter at 0x11851c730>], 'vect__ngram_range': [(1, 1)], 'vect__stop_words': [['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', '...alty': ['l1', 'l2'], 'vect__norm': [None], 'vect__ngram_range': [(1, 1)], 'vect__use_idf': [False]}],\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring='accuracy', verbose=1)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gs_lr_tfidf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best parameter set: {'vect__tokenizer': <function tokenizer at 0x11851c6a8>, 'clf__C': 10.0, 'vect__stop_words': None, 'clf__penalty': 'l2', 'vect__ngram_range': (1, 1)} \n",
      "CV Accuracy: 0.897\n"
     ]
    }
   ],
   "source": [
    "print('Best parameter set: %s ' % gs_lr_tfidf.best_params_)\n",
    "print('CV Accuracy: %.3f' % gs_lr_tfidf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy: 0.899\n"
     ]
    }
   ],
   "source": [
    "clf = gs_lr_tfidf.best_estimator_\n",
    "print('Test Accuracy: %.3f' % clf.score(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "<hr>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Start comment:\n",
    "    \n",
    "Please note that `gs_lr_tfidf.best_score_` is the average k-fold cross-validation score. I.e., if we have a `GridSearchCV` object with 5-fold cross-validation (like the one above), the `best_score_` attribute returns the average score over the 5-folds of the best model. To illustrate this with an example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.6,  0.4,  0.6,  0.2,  0.6])"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "import numpy as np\n",
    "if Version(sklearn_version) < '0.18':\n",
    "    from sklearn.cross_validation import StratifiedKFold\n",
    "    from sklearn.cross_validation import cross_val_score\n",
    "else:\n",
    "    from sklearn.model_selection import StratifiedKFold\n",
    "    from sklearn.model_selection import cross_val_score\n",
    "\n",
    "np.random.seed(0)\n",
    "np.set_printoptions(precision=6)\n",
    "y = [np.random.randint(3) for i in range(25)]\n",
    "X = (y + np.random.randn(25)).reshape(-1, 1)\n",
    "\n",
    "if Version(sklearn_version) < '0.18':\n",
    "    cv5_idx = list(StratifiedKFold(y, n_folds=5, shuffle=False, random_state=0))\n",
    "\n",
    "else:\n",
    "    cv5_idx = list(StratifiedKFold(n_splits=5, shuffle=False, random_state=0).split(X, y))\n",
    "    \n",
    "cross_val_score(LogisticRegression(random_state=123), X, y, cv=cv5_idx)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By executing the code above, we created a simple data set of random integers that shall represent our class labels. Next, we fed the indices of 5 cross-validation folds (`cv3_idx`) to the `cross_val_score` scorer, which returned 5 accuracy scores -- these are the 5 accuracy values for the 5 test folds.  \n",
    "\n",
    "Next, let us use the `GridSearchCV` object and feed it the same 5 cross-validation sets (via the pre-generated `cv3_idx` indices):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
      "[CV]  ................................................................\n",
      "[CV] ....................................... , score=0.600000 -   0.0s\n",
      "[CV]  ................................................................\n",
      "[CV] ....................................... , score=0.400000 -   0.0s\n",
      "[CV]  ................................................................\n",
      "[CV] ....................................... , score=0.600000 -   0.0s\n",
      "[CV]  ................................................................\n",
      "[CV] ....................................... , score=0.200000 -   0.0s\n",
      "[CV]  ................................................................\n",
      "[CV] ....................................... , score=0.600000 -   0.0s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   5 out of   5 | elapsed:    0.0s finished\n"
     ]
    }
   ],
   "source": [
    "if Version(sklearn_version) < '0.18':\n",
    "    from sklearn.grid_search import GridSearchCV\n",
    "else:\n",
    "    from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "gs = GridSearchCV(LogisticRegression(), {}, cv=cv5_idx, verbose=3).fit(X, y) \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, the scores for the 5 folds are exactly the same as the ones from `cross_val_score` earlier."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, the best_score_ attribute of the `GridSearchCV` object, which becomes available after `fit`ting, returns the average accuracy score of the best model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.47999999999999998"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gs.best_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, the result above is consistent with the average score computed the `cross_val_score`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.47999999999999998"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cross_val_score(LogisticRegression(), X, y, cv=cv5_idx).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### End comment.\n",
    "\n",
    "<hr>\n",
    "<hr>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Working with bigger data - online algorithms and out-of-core learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import re\n",
    "from nltk.corpus import stopwords\n",
    "\n",
    "def tokenizer(text):\n",
    "    text = re.sub('<[^>]*>', '', text)\n",
    "    emoticons = re.findall('(?::|;|=)(?:-)?(?:\\)|\\(|D|P)', text.lower())\n",
    "    text = re.sub('[\\W]+', ' ', text.lower()) +\\\n",
    "        ' '.join(emoticons).replace('-', '')\n",
    "    tokenized = [w for w in text.split() if w not in stop]\n",
    "    return tokenized\n",
    "\n",
    "\n",
    "def stream_docs(path):\n",
    "    with open(path, 'r', encoding='utf-8') as csv:\n",
    "        next(csv)  # skip header\n",
    "        for line in csv:\n",
    "            text, label = line[:-3], int(line[-2])\n",
    "            yield text, label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('\"In 1974, the teenager Martha Moxley (Maggie Grace) moves to the high-class area of Belle Haven, Greenwich, Connecticut. On the Mischief Night, eve of Halloween, she was murdered in the backyard of her house and her murder remained unsolved. Twenty-two years later, the writer Mark Fuhrman (Christopher Meloni), who is a former LA detective that has fallen in disgrace for perjury in O.J. Simpson trial and moved to Idaho, decides to investigate the case with his partner Stephen Weeks (Andrew Mitchell) with the purpose of writing a book. The locals squirm and do not welcome them, but with the support of the retired detective Steve Carroll (Robert Forster) that was in charge of the investigation in the 70\\'s, they discover the criminal and a net of power and money to cover the murder.<br /><br />\"\"Murder in Greenwich\"\" is a good TV movie, with the true story of a murder of a fifteen years old girl that was committed by a wealthy teenager whose mother was a Kennedy. The powerful and rich family used their influence to cover the murder for more than twenty years. However, a snoopy detective and convicted perjurer in disgrace was able to disclose how the hideous crime was committed. The screenplay shows the investigation of Mark and the last days of Martha in parallel, but there is a lack of the emotion in the dramatization. My vote is seven.<br /><br />Title (Brazil): Not Available\"',\n",
       " 1)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "next(stream_docs(path='./movie_data.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def get_minibatch(doc_stream, size):\n",
    "    docs, y = [], []\n",
    "    try:\n",
    "        for _ in range(size):\n",
    "            text, label = next(doc_stream)\n",
    "            docs.append(text)\n",
    "            y.append(label)\n",
    "    except StopIteration:\n",
    "        return None, None\n",
    "    return docs, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import HashingVectorizer\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "\n",
    "vect = HashingVectorizer(decode_error='ignore', \n",
    "                         n_features=2**21,\n",
    "                         preprocessor=None, \n",
    "                         tokenizer=tokenizer)\n",
    "\n",
    "clf = SGDClassifier(loss='log', random_state=1, n_iter=1)\n",
    "doc_stream = stream_docs(path='./movie_data.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "0%                          100%\n",
      "[##############################] | ETA: 00:00:00\n",
      "Total time elapsed: 00:00:44\n"
     ]
    }
   ],
   "source": [
    "import pyprind\n",
    "pbar = pyprind.ProgBar(45)\n",
    "\n",
    "classes = np.array([0, 1])\n",
    "for _ in range(45):\n",
    "    X_train, y_train = get_minibatch(doc_stream, size=1000)\n",
    "    if not X_train:\n",
    "        break\n",
    "    X_train = vect.transform(X_train)\n",
    "    clf.partial_fit(X_train, y_train, classes=classes)\n",
    "    pbar.update()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.867\n"
     ]
    }
   ],
   "source": [
    "X_test, y_test = get_minibatch(doc_stream, size=5000)\n",
    "X_test = vect.transform(X_test)\n",
    "print('Accuracy: %.3f' % clf.score(X_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "clf = clf.partial_fit(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Summary"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}