chapter08_ml/04_text.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 8.4. Learning from text — Naive Bayes for Natural Language Processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import sklearn\n",
    "import sklearn.model_selection as ms\n",
    "import sklearn.feature_extraction.text as text\n",
    "import sklearn.naive_bayes as nb\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('https://github.com/ipython-books/'\n",
    "                 'cookbook-2nd-data/blob/master/'\n",
    "                 'troll.csv?raw=true')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "podoc": {
     "output_text": "Output"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhEAAAGCEAAAAAArfxL9AAAJJGlDQ1BpY2MAAHjalZVnUJNZF8fv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\nSUVORK5CYII=\n",
      "text/plain": [
       "Output"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df[['Insult', 'Comment']].tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = df['Insult']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3947, 16469)\n"
     ]
    }
   ],
   "source": [
    "tf = text.TfidfVectorizer()\n",
    "X = tf.fit_transform(df['Comment'])\n",
    "print(X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Each sample has ~0.15% non-zero features.\n"
     ]
    }
   ],
   "source": [
    "p = 100 * X.nnz / float(X.shape[0] * X.shape[1])\n",
    "print(f\"Each sample has ~{p:.2f}% non-zero features.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "(X_train, X_test, y_train, y_test) = \\\n",
    "    ms.train_test_split(X, y, test_size=.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "bnb = ms.GridSearchCV(\n",
    "    nb.BernoulliNB(),\n",
    "    param_grid={'alpha': np.logspace(-2., 2., 50)})\n",
    "bnb.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.761"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bnb.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "you,are,your,to,the,and,of,that,is,in,it,like,have,on,not,for,just,re,with,be,an,so,this,xa0,all,idiot,what,get,up,go,****,don,stupid,no,as,do,can,***,or,but,if,know,who,about,dumb,****,me,******,because,back\n"
     ]
    }
   ],
   "source": [
    "# We first get the words corresponding to each feature\n",
    "names = np.asarray(tf.get_feature_names())\n",
    "# Next, we display the 50 words with the largest\n",
    "# coefficients.\n",
    "print(','.join(names[np.argsort(\n",
    "    bnb.best_estimator_.coef_[0, :])[::-1][:50]]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 1]\n"
     ]
    }
   ],
   "source": [
    "print(bnb.predict(tf.transform([\n",
    "    \"I totally agree with you.\",\n",
    "    \"You are so stupid.\"\n",
    "])))"
   ]
  }
 ],
 "metadata": {},
 "nbformat": 4,
 "nbformat_minor": 2
}