project.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Spam sms detection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Importing packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import nltk\n",
    "import re\n",
    "import string\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer\n",
    "\n",
    "stopwords=nltk.corpus.stopwords.words('english')\n",
    "ps=nltk.PorterStemmer()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data reading"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "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>label</th>\n",
       "      <th>body_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>spam</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ham</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ham</td>\n",
       "      <td>Even my brother is not like to speak with me. ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  label                                          body_text\n",
       "0  spam  Free entry in 2 a wkly comp to win FA Cup fina...\n",
       "1   ham  Nah I don't think he goes to usf, he lives aro...\n",
       "2   ham  Even my brother is not like to speak with me. ..."
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv(\"./data/SMSSpamCollection.tsv\", sep='\\t')\n",
    "data.columns = ['label', 'body_text']\n",
    "data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5567, 2)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "def count_punct(text):\n",
    "    count=sum([1 for char in text if char in string.punctuation])\n",
    "    return round(count/(len(text)-text.count(' ')),3) *100\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "data['body_length']=data['body_text'].apply(lambda x: len(x) - x.count(' '))\n",
    "data['punct%']= data['body_text'].apply(lambda x: count_punct(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib import pyplot\n",
    "import numpy as np\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins = np.linspace(0, 200, 40)\n",
    "\n",
    "pyplot.hist(data[data['label']=='spam']['body_length'], bins, alpha=0.5, density=True, label='spam')\n",
    "pyplot.hist(data[data['label']=='ham']['body_length'], bins, alpha=0.5, density=True, label='ham')\n",
    "pyplot.legend(loc='upper left')\n",
    "pyplot.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins = np.linspace(0, 50, 40)\n",
    "\n",
    "pyplot.hist(data[data['label']=='spam']['punct%'], bins, alpha=0.5, density=True, label='spam')\n",
    "pyplot.hist(data[data['label']=='ham']['punct%'], bins, alpha=0.5, density=True, label='ham')\n",
    "pyplot.legend(loc='upper right')\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### New features transformation. Box-Cox Power Transformation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins = np.linspace(0, 200, 40)\n",
    "\n",
    "pyplot.hist(data['body_length'], bins)\n",
    "pyplot.title(\"Body Length Distribution\")\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins = np.linspace(0, 50, 40)\n",
    "\n",
    "pyplot.hist(data['punct%'], bins)\n",
    "pyplot.title(\"Punctuation % Distribution\")\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAW6klEQVR4nO3df5Bd5X3f8fcnwohfJQizECEpWblVHAsax7ZGIXbqupUTxA8jZjKkco2tuHTUeGhtZ/A4UsiEOq06Suom2G3FRMWAbBgUBTtBMcWGKCE0DQYvP2IQQka2sLSWQIsJBttURuLTP+6j9rDcXe3eu3vXus/nNbNzz/me59zzPIv43LPPOfde2SYiIurwYzPdgYiI6J2EfkRERRL6EREVSehHRFQkoR8RUZGEfkRERRL6EREVSehH35F0lqR7Jb0o6b/MdH9Gk/Rbkq6f6X5EnRL6MaUkfa/x84qklxrr7+tRN1YDzwKn2r6qR8dsS9K7JA03a7b/k+1/3YNjz5W0VdI+SZY02KbN8ZKelXSKpF+V9LeSfiDpnunuX8yMhH5MKdunHPkB9gDvadRuOdJO0nHT2I2fAh53B283n+Z+9dorwJeAXxmnzTuBR2x/D3gOuBZY34O+xQxJ6EdPHDnjlfSbkp4GbpQ0R9IXJY1I+vuyPL+xzz2S/oOk/12mau6SdEbZdoKkmyV9R9Lzkr5apnVuAlYBHy9/Xbxb0mxJ15Yz3n1lefY4/fr3kv6kPP+Lkh6V9NOS1ko6IGmvpF9u9PODknaUtt+U9G9K/WTgTuDsxl87Z5fnv7mx/yWStpdx3CPpTY1tT0n6mKSvSfqupD+WdMJEfue2n7G9AfjqOM0uBP5naf8XtrcA+yby/HFsSuhHL/0EcDqtM/HVtP793VjWfxJ4Cfhvo/b5l8AHgTOB44GPlfoq4MeBBcDrgV8HXrL9a8AtwO+Xvy7+ArgaOA/4OeDNwFLgt8fpF8B7gM8Bc4CHgS+X/s4Dfhf4o8b+B4CLgVNLX/9Q0lttfx+4ANjX+GvnVYEq6aeBW4GPAgO0AvjPJR3faParwHJgIfCzwK819n9e0i/SuQuBO7rYP44xCf3opVeAa2wftP2S7e/Y/rztH9h+EVgH/NNR+9xo++u2XwK20ApugJdphf0/sn3Y9oO2XxjjuO8Dftf2AdsjwCeA94/Vr1L7X7a/bPsQ8Ce0Anm97ZeBzcCgpNMAbN9h+xtu+WvgLuCfTPB38i+AO2zfXZ77k8CJwNsbbT5te5/t54A/b/wOsH2a7b+Z4LFeRdIbgNfZ3tnJ/nFsSuhHL43Y/j9HViSdJOmPJH1L0gvAvcBpkmY19nm6sfwD4JSy/DlaZ9+by5TN70t63RjHPRv4VmP9W6XWtl/FM43ll4BnbR9urHOkL5IukPQVSc9Jep7W2fMZY/Rl3L7ZfgXYS+sviiPG+h106yLK1E7UI6EfvTT6wupVwBuBn7d9Kq2LigA66hPZL9v+hO3FtM6KLwY+MEbzfbSmbo74SV49b93x54uXawOfp3WGfpbt02gF6ZExHO25X9U3SaI1ZfXtTvs0CZnaqVBCP2bSP6B11vy8pNOBaya6o6R/Jukfl78KXqA13XN4jOa3Ar8taaBcCP4d4OYx2k7W8cBsYAQ4JOkC4Jcb258BXi/px8fYfwtwkaRl5S+Vq4CDwN9ORefKRd/ZZXX2kYvAkk6kdW3jnkbbWWX7ccCPlYvlY/31FMeohH7MpGtpzV8/C3yF1u2FE/UTwG20An8H8NeMHeT/ERgCvgY8CjxUal0r1yI+TCu8/57Wheetje1P0HrR+Wa56Hr2qP13ApcD/5XW7+E9tG5z/eFEjl/uCBrv+sFLwPfK8hP8/6mpZcB9o6a13l+2X0frmsRLwP+YSD/i2KF8c1ZEfSRtAB4rt3RGRfrpjSgRMXGP0LoTKCqTM/2IiIpkTj8ioiI/8tM7Z5xxhgcHB2e6GxERx5QHH3zwWdsDo+s/8qE/ODjI0NDQTHcjIuKYIulb7eqZ3omIqEhCPyKiIgn9iIiKJPQjIiqS0I+IqEhCPyKiIgn9iIiKJPQjIiqS0I+IqMiP/Dtyp9PgmrG/NOip9Rf1sCcREb2RM/2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioyFFDX9INkg5IeqzNto9JsqQzGrW1knZJ2inp/Eb9bZIeLds+LUlTN4yIiJiIiZzp3wQsH12UtAD4JWBPo7YYWAmcU/bZIGlW2XwdsBpYVH5e85wRETG9jhr6tu8Fnmuz6Q+BjwNu1FYAm20ftL0b2AUslTQXONX2fbYNfBa4tOveR0TEpHQ0py/pEuDbtv9u1KZ5wN7G+nCpzSvLo+tjPf9qSUOShkZGRjrpYkREtDHp0Jd0EnA18DvtNrepeZx6W7Y32l5ie8nAwMBkuxgREWPo5EtU/iGwEPi7ci12PvCQpKW0zuAXNNrOB/aV+vw29YiI6KFJn+nbftT2mbYHbQ/SCvS32n4a2AqslDRb0kJaF2wfsL0feFHSeeWunQ8At0/dMCIiYiImcsvmrcB9wBslDUu6Yqy2trcDW4DHgS8BV9o+XDZ/CLie1sXdbwB3dtn3iIiYpKNO79h+71G2D45aXwesa9NuCDh3kv2LiIgplHfkRkRUJKEfEVGRhH5EREUS+hERFUnoR0RUJKEfEVGRhH5EREUS+hERFUnoR0RUJKEfEVGRhH5EREUS+hERFUnoR0RUJKEfEVGRhH5EREU6+brEKgyuuWPc7U+tv6hHPYmImDo504+IqEhCPyKiIhP5jtwbJB2Q9Fij9p8lPSHpa5L+VNJpjW1rJe2StFPS+Y362yQ9WrZ9unxBekRE9NBEzvRvApaPqt0NnGv7Z4GvA2sBJC0GVgLnlH02SJpV9rkOWA0sKj+jnzMiIqbZUUPf9r3Ac6Nqd9k+VFa/AswvyyuAzbYP2t4N7AKWSpoLnGr7PtsGPgtcOlWDiIiIiZmKOf1/BdxZlucBexvbhkttXlkeXY+IiB7qKvQlXQ0cAm45UmrTzOPUx3re1ZKGJA2NjIx008WIiGjoOPQlrQIuBt5XpmygdQa/oNFsPrCv1Oe3qbdle6PtJbaXDAwMdNrFiIgYpaPQl7Qc+E3gEts/aGzaCqyUNFvSQloXbB+wvR94UdJ55a6dDwC3d9n3iIiYpKO+I1fSrcC7gDMkDQPX0LpbZzZwd7nz8iu2f932dklbgMdpTftcaftweaoP0boT6ERa1wDuJCIieuqooW/7vW3Knxmn/TpgXZv6EHDupHoXERFTKu/IjYioSEI/IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIocNfQl3SDpgKTHGrXTJd0t6cnyOKexba2kXZJ2Sjq/UX+bpEfLtk+rfKN6RET0zkTO9G8Clo+qrQG22V4EbCvrSFoMrATOKftskDSr7HMdsBpYVH5GP2dEREyzo4a+7XuB50aVVwCbyvIm4NJGfbPtg7Z3A7uApZLmAqfavs+2gc829omIiB7pdE7/LNv7AcrjmaU+D9jbaDdcavPK8uh6W5JWSxqSNDQyMtJhFyMiYrSpvpDbbp7e49Tbsr3R9hLbSwYGBqascxERtes09J8pUzaUxwOlPgwsaLSbD+wr9flt6hER0UOdhv5WYFVZXgXc3qivlDRb0kJaF2wfKFNAL0o6r9y184HGPhER0SPHHa2BpFuBdwFnSBoGrgHWA1skXQHsAS4DsL1d0hbgceAQcKXtw+WpPkTrTqATgTvLT0RE9NBRQ9/2e8fYtGyM9uuAdW3qQ8C5k+pdRERMqbwjNyKiIgn9iIiKJPQjIiqS0I+IqEhCPyKiIke9e+dYNrjmjpnuQkTEj5Sc6UdEVCShHxFRkYR+RERFEvoRERVJ6EdEVCShHxFRkYR+RERFEvoRERVJ6EdEVCShHxFRkYR+RERFEvoRERXpKvQl/Yak7ZIek3SrpBMknS7pbklPlsc5jfZrJe2StFPS+d13PyIiJqPj0Jc0D/gwsMT2ucAsYCWwBthmexGwrawjaXHZfg6wHNggaVZ33Y+IiMnodnrnOOBESccBJwH7gBXAprJ9E3BpWV4BbLZ90PZuYBewtMvjR0TEJHQc+ra/DXwS2APsB75r+y7gLNv7S5v9wJlll3nA3sZTDJfaa0haLWlI0tDIyEinXYyIiFG6md6ZQ+vsfSFwNnCypMvH26VNze0a2t5oe4ntJQMDA512MSIiRulmeufdwG7bI7ZfBr4AvB14RtJcgPJ4oLQfBhY09p9PazooIiJ6pJvQ3wOcJ+kkSQKWATuArcCq0mYVcHtZ3gqslDRb0kJgEfBAF8ePiIhJ6vg7cm3fL+k24CHgEPAwsBE4Bdgi6QpaLwyXlfbbJW0BHi/tr7R9uMv+R0TEJHT1xei2rwGuGVU+SOusv137dcC6bo4ZERGdyztyIyIqktCPiKhIQj8ioiIJ/YiIiiT0IyIqktCPiKhIQj8ioiIJ/YiIiiT0IyIq0tU7cms2uOaOMbc9tf6iHvYkImLicqYfEVGRhH5EREUS+hERFUnoR0RUJKEfEVGRhH5EREUS+hERFUnoR0RUJKEfEVGRrkJf0mmSbpP0hKQdkn5B0umS7pb0ZHmc02i/VtIuSTslnd999yMiYjK6PdP/FPAl2z8DvBnYAawBttleBGwr60haDKwEzgGWAxskzery+BERMQkdh76kU4F3Ap8BsP1D288DK4BNpdkm4NKyvALYbPug7d3ALmBpp8ePiIjJ6+ZM/w3ACHCjpIclXS/pZOAs2/sByuOZpf08YG9j/+FSew1JqyUNSRoaGRnpoosREdHUTegfB7wVuM72W4DvU6ZyxqA2NbdraHuj7SW2lwwMDHTRxYiIaOom9IeBYdv3l/XbaL0IPCNpLkB5PNBov6Cx/3xgXxfHj4iISeo49G0/DeyV9MZSWgY8DmwFVpXaKuD2srwVWClptqSFwCLggU6PHxERk9ftl6j8O+AWSccD3wQ+SOuFZIukK4A9wGUAtrdL2kLrheEQcKXtw10ePyIiJqGr0Lf9CLCkzaZlY7RfB6zr5pgREdG5vCM3IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioSEI/IqIiCf2IiIok9CMiKpLQj4ioSLffnBVtDK65Y9ztT62/qEc9iYh4tZzpR0RUpOvQlzRL0sOSvljWT5d0t6Qny+OcRtu1knZJ2inp/G6PHRERkzMVZ/ofAXY01tcA22wvAraVdSQtBlYC5wDLgQ2SZk3B8SMiYoK6Cn1J84GLgOsb5RXAprK8Cbi0Ud9s+6Dt3cAuYGk3x4+IiMnp9kz/WuDjwCuN2lm29wOUxzNLfR6wt9FuuNReQ9JqSUOShkZGRrrsYkREHNFx6Eu6GDhg+8GJ7tKm5nYNbW+0vcT2koGBgU67GBERo3Rzy+Y7gEskXQicAJwq6WbgGUlzbe+XNBc4UNoPAwsa+88H9nVx/IiImKSOz/Rtr7U93/YgrQu0f2n7cmArsKo0WwXcXpa3AislzZa0EFgEPNBxzyMiYtKm481Z64Etkq4A9gCXAdjeLmkL8DhwCLjS9uFpOH5ERIxhSkLf9j3APWX5O8CyMdqtA9ZNxTEjImLy8o7ciIiKJPQjIiqS0I+IqEhCPyKiIgn9iIiKJPQjIiqS0I+IqEhCPyKiIgn9iIiKJPQjIiqS0I+IqEhCPyKiIgn9iIiKJPQjIioyHZ+nH0cxuOaOcbc/tf6iHvUkImqTM/2IiIok9CMiKpLQj4ioSMehL2mBpL+StEPSdkkfKfXTJd0t6cnyOKexz1pJuyTtlHT+VAwgIiImrpsz/UPAVbbfBJwHXClpMbAG2GZ7EbCtrFO2rQTOAZYDGyTN6qbzERExOR2Hvu39th8qyy8CO4B5wApgU2m2Cbi0LK8ANts+aHs3sAtY2unxIyJi8qZkTl/SIPAW4H7gLNv7ofXCAJxZms0D9jZ2Gy61ds+3WtKQpKGRkZGp6GJERDAFoS/pFODzwEdtvzBe0zY1t2toe6PtJbaXDAwMdNvFiIgounpzlqTX0Qr8W2x/oZSfkTTX9n5Jc4EDpT4MLGjsPh/Y183x+9V4b97KG7ciohvd3L0j4DPADtt/0Ni0FVhVllcBtzfqKyXNlrQQWAQ80OnxIyJi8ro5038H8H7gUUmPlNpvAeuBLZKuAPYAlwHY3i5pC/A4rTt/rrR9uIvjR0TEJHUc+rb/hvbz9ADLxthnHbCu02NGRER38o7ciIiKJPQjIiqS0I+IqEhCPyKiIgn9iIiKJPQjIiqS0I+IqEhCPyKiIgn9iIiKJPQjIirS1adsRu+N9wmckE/hjIjx5Uw/IqIiCf2IiIpkeqfP5AtYImI8OdOPiKhIQj8ioiKZ3qlI7vyJiIR+/D95UYjof5neiYioSM/P9CUtBz4FzAKut72+132IqZe/EiKODT0NfUmzgP8O/BIwDHxV0lbbj/eyH9GZowV7p/vmBWHy8vuMTvX6TH8psMv2NwEkbQZWAAn9inXzYtKtbgIyf93EsajXoT8P2NtYHwZ+fnQjSauB1WX1e5J2dni8M4BnO9z3WJUxT4J+b4p70qPnZpwxT/NxZ0r+XU/eT7Ur9jr01abm1xTsjcDGrg8mDdle0u3zHEsy5jrUNubaxgvTN+Ze370zDCxorM8H9vW4DxER1ep16H8VWCRpoaTjgZXA1h73ISKiWj2d3rF9SNK/Bb5M65bNG2xvn8ZDdj1FdAzKmOtQ25hrGy9M05hlv2ZKPSIi+lTekRsRUZGEfkRERfoy9CUtl7RT0i5Ja2a6P9NB0gJJfyVph6Ttkj5S6qdLulvSk+Vxzkz3dapJmiXpYUlfLOt9PWZJp0m6TdIT5b/3L1Qw5t8o/64fk3SrpBP6bcySbpB0QNJjjdqYY5S0tmTaTknnd3rcvgv9xkc9XAAsBt4rafHM9mpaHAKusv0m4DzgyjLONcA224uAbWW933wE2NFY7/cxfwr4ku2fAd5Ma+x9O2ZJ84APA0tsn0vrpo+V9N+YbwKWj6q1HWP5f3slcE7ZZ0PJuknru9Cn8VEPtn8IHPmoh75ie7/th8ryi7SCYB6tsW4qzTYBl85MD6eHpPnARcD1jXLfjlnSqcA7gc8A2P6h7efp4zEXxwEnSjoOOInW+3n6asy27wWeG1Uea4wrgM22D9reDeyilXWT1o+h3+6jHubNUF96QtIg8BbgfuAs2/uh9cIAnDlzPZsW1wIfB15p1Pp5zG8ARoAby5TW9ZJOpo/HbPvbwCeBPcB+4Lu276KPx9ww1hinLNf6MfQn9FEP/ULSKcDngY/afmGm+zOdJF0MHLD94Ez3pYeOA94KXGf7LcD3OfanNcZV5rFXAAuBs4GTJV0+s72acVOWa/0Y+tV81IOk19EK/Ftsf6GUn5E0t2yfCxyYqf5Ng3cAl0h6ita03T+XdDP9PeZhYNj2/WX9NlovAv085ncDu22P2H4Z+ALwdvp7zEeMNcYpy7V+DP0qPupBkmjN8+6w/QeNTVuBVWV5FXB7r/s2XWyvtT3f9iCt/65/afty+nvMTwN7Jb2xlJbR+ijyvh0zrWmd8ySdVP6dL6N1zaqfx3zEWGPcCqyUNFvSQmAR8EBHR7Dddz/AhcDXgW8AV890f6ZpjL9I68+7rwGPlJ8LgdfTuur/ZHk8fab7Ok3jfxfwxbLc12MGfg4YKv+t/wyYU8GYPwE8ATwGfA6Y3W9jBm6ldc3iZVpn8leMN0bg6pJpO4ELOj1uPoYhIqIi/Ti9ExERY0joR0RUJKEfEVGRhH5EREUS+hERFUnoR0RUJKEfEVGR/wsraMO4UqpKBAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in [1,2,3,4,5]:\n",
    "    pyplot.hist((data['punct%'])**(1/i), bins=40)\n",
    "    pyplot.title(\"Transformation: 1/{}\".format(str(i)))\n",
    "    pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "def data_clean(text):\n",
    "    text=''.join([word.lower() for word in text if word not in string.punctuation])\n",
    "    tokens= re.split('\\W+', text)\n",
    "    text=[ps.stem(word) for word in tokens if word not in stopwords]\n",
    "    return text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Vectorizing TF_IDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "tfidf_vect=TfidfVectorizer(analyzer=data_clean)\n",
    "X_tfidf=tfidf_vect.fit_transform(data['body_text'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "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>body_length</th>\n",
       "      <th>punct%</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>...</th>\n",
       "      <th>8094</th>\n",
       "      <th>8095</th>\n",
       "      <th>8096</th>\n",
       "      <th>8097</th>\n",
       "      <th>8098</th>\n",
       "      <th>8099</th>\n",
       "      <th>8100</th>\n",
       "      <th>8101</th>\n",
       "      <th>8102</th>\n",
       "      <th>8103</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>128</td>\n",
       "      <td>4.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>49</td>\n",
       "      <td>4.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>62</td>\n",
       "      <td>3.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 8106 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   body_length  punct%    0    1    2    3    4    5    6    7  ...  8094  \\\n",
       "0          128     4.7  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   0.0   \n",
       "1           49     4.1  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   0.0   \n",
       "2           62     3.2  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...   0.0   \n",
       "\n",
       "   8095  8096  8097  8098  8099  8100  8101  8102  8103  \n",
       "0   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0  \n",
       "1   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0  \n",
       "2   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0  \n",
       "\n",
       "[3 rows x 8106 columns]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_features=pd.concat([data['body_length'], data['punct%'], pd.DataFrame(X_tfidf.toarray())], axis=1)\n",
    "X_features.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Vectorizing CountVectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "count_vect=CountVectorizer(analyzer=data_clean)\n",
    "X_count=count_vect.fit_transform(data['body_text'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "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>body_length</th>\n",
       "      <th>punct%</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>...</th>\n",
       "      <th>8094</th>\n",
       "      <th>8095</th>\n",
       "      <th>8096</th>\n",
       "      <th>8097</th>\n",
       "      <th>8098</th>\n",
       "      <th>8099</th>\n",
       "      <th>8100</th>\n",
       "      <th>8101</th>\n",
       "      <th>8102</th>\n",
       "      <th>8103</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>128</td>\n",
       "      <td>4.7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>49</td>\n",
       "      <td>4.1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>62</td>\n",
       "      <td>3.2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 8106 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   body_length  punct%  0  1  2  3  4  5  6  7  ...  8094  8095  8096  8097  \\\n",
       "0          128     4.7  0  0  0  0  0  0  0  0  ...     0     0     0     0   \n",
       "1           49     4.1  0  0  0  0  0  0  0  0  ...     0     0     0     0   \n",
       "2           62     3.2  0  0  0  0  0  0  0  0  ...     0     0     0     0   \n",
       "\n",
       "   8098  8099  8100  8101  8102  8103  \n",
       "0     0     0     0     0     0     0  \n",
       "1     0     0     0     0     0     0  \n",
       "2     0     0     0     0     0     0  \n",
       "\n",
       "[3 rows x 8106 columns]"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_count_features=pd.concat([data['body_length'], data['punct%'], pd.DataFrame(X_count.toarray())], axis=1)\n",
    "X_count_features.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Building Machine Learning Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import KFold, cross_val_score\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import GradientBoostingClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest cv_score: 0.9737726654633263\n"
     ]
    }
   ],
   "source": [
    "rf=RandomForestClassifier(n_jobs=-1)\n",
    "K_Fold=KFold(n_splits=5)\n",
    "cv_score=cross_val_score(rf, X_features, data['label'], cv=K_Fold, scoring='accuracy', n_jobs=-1)\n",
    "print('Random Forest cv_score:', cv_score.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GRadient Boost cv_score: 0.9615569868745574\n"
     ]
    }
   ],
   "source": [
    "gb=GradientBoostingClassifier()\n",
    "K_Fold=KFold(n_splits=5)\n",
    "cv_score=cross_val_score(gb, X_features, data['label'], cv=K_Fold, scoring='accuracy', n_jobs=-1)\n",
    "print('GRadient Boost cv_score:', cv_score.mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Tuning parameters setting using GridSearch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\", category=DeprecationWarning)\n",
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>3.282062</td>\n",
       "      <td>0.035623</td>\n",
       "      <td>0.224661</td>\n",
       "      <td>0.014933</td>\n",
       "      <td>90</td>\n",
       "      <td>10</td>\n",
       "      <td>{'max_depth': 90, 'n_estimators': 10}</td>\n",
       "      <td>0.975763</td>\n",
       "      <td>0.978456</td>\n",
       "      <td>0.978437</td>\n",
       "      <td>0.966757</td>\n",
       "      <td>0.977538</td>\n",
       "      <td>0.975390</td>\n",
       "      <td>0.004427</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>40.996854</td>\n",
       "      <td>1.324907</td>\n",
       "      <td>0.424428</td>\n",
       "      <td>0.080756</td>\n",
       "      <td>90</td>\n",
       "      <td>300</td>\n",
       "      <td>{'max_depth': 90, 'n_estimators': 300}</td>\n",
       "      <td>0.977558</td>\n",
       "      <td>0.978456</td>\n",
       "      <td>0.974843</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.970350</td>\n",
       "      <td>0.973952</td>\n",
       "      <td>0.003904</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>23.882267</td>\n",
       "      <td>0.282269</td>\n",
       "      <td>0.362860</td>\n",
       "      <td>0.037763</td>\n",
       "      <td>None</td>\n",
       "      <td>150</td>\n",
       "      <td>{'max_depth': None, 'n_estimators': 150}</td>\n",
       "      <td>0.977558</td>\n",
       "      <td>0.972172</td>\n",
       "      <td>0.974843</td>\n",
       "      <td>0.966757</td>\n",
       "      <td>0.975741</td>\n",
       "      <td>0.973414</td>\n",
       "      <td>0.003755</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>33.444902</td>\n",
       "      <td>3.949127</td>\n",
       "      <td>0.286280</td>\n",
       "      <td>0.022408</td>\n",
       "      <td>None</td>\n",
       "      <td>300</td>\n",
       "      <td>{'max_depth': None, 'n_estimators': 300}</td>\n",
       "      <td>0.978456</td>\n",
       "      <td>0.977558</td>\n",
       "      <td>0.972147</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.969452</td>\n",
       "      <td>0.973233</td>\n",
       "      <td>0.004083</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.142819</td>\n",
       "      <td>0.171635</td>\n",
       "      <td>0.233567</td>\n",
       "      <td>0.017639</td>\n",
       "      <td>60</td>\n",
       "      <td>10</td>\n",
       "      <td>{'max_depth': 60, 'n_estimators': 10}</td>\n",
       "      <td>0.980251</td>\n",
       "      <td>0.976661</td>\n",
       "      <td>0.973046</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.965858</td>\n",
       "      <td>0.972874</td>\n",
       "      <td>0.005228</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "6        3.282062      0.035623         0.224661        0.014933   \n",
       "8       40.996854      1.324907         0.424428        0.080756   \n",
       "10      23.882267      0.282269         0.362860        0.037763   \n",
       "11      33.444902      3.949127         0.286280        0.022408   \n",
       "3        3.142819      0.171635         0.233567        0.017639   \n",
       "\n",
       "   param_max_depth param_n_estimators  \\\n",
       "6               90                 10   \n",
       "8               90                300   \n",
       "10            None                150   \n",
       "11            None                300   \n",
       "3               60                 10   \n",
       "\n",
       "                                      params  split0_test_score  \\\n",
       "6      {'max_depth': 90, 'n_estimators': 10}           0.975763   \n",
       "8     {'max_depth': 90, 'n_estimators': 300}           0.977558   \n",
       "10  {'max_depth': None, 'n_estimators': 150}           0.977558   \n",
       "11  {'max_depth': None, 'n_estimators': 300}           0.978456   \n",
       "3      {'max_depth': 60, 'n_estimators': 10}           0.980251   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "6            0.978456           0.978437           0.966757   \n",
       "8            0.978456           0.974843           0.968553   \n",
       "10           0.972172           0.974843           0.966757   \n",
       "11           0.977558           0.972147           0.968553   \n",
       "3            0.976661           0.973046           0.968553   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \n",
       "6            0.977538         0.975390        0.004427                1  \n",
       "8            0.970350         0.973952        0.003904                2  \n",
       "10           0.975741         0.973414        0.003755                3  \n",
       "11           0.969452         0.973233        0.004083                4  \n",
       "3            0.965858         0.972874        0.005228                5  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestClassifier()\n",
    "param = {'n_estimators': [10, 150, 300],\n",
    "        'max_depth': [30, 60, 90, None]}\n",
    "gs = GridSearchCV(rf, param, cv=5, n_jobs=-1)\n",
    "gs_fit = gs.fit(X_features, data['label'])\n",
    "pd.DataFrame(gs_fit.cv_results_).sort_values('mean_test_score', ascending=False)[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>21.955820</td>\n",
       "      <td>0.180804</td>\n",
       "      <td>0.339013</td>\n",
       "      <td>0.007180</td>\n",
       "      <td>90</td>\n",
       "      <td>150</td>\n",
       "      <td>{'max_depth': 90, 'n_estimators': 150}</td>\n",
       "      <td>0.975763</td>\n",
       "      <td>0.975763</td>\n",
       "      <td>0.973944</td>\n",
       "      <td>0.966757</td>\n",
       "      <td>0.973944</td>\n",
       "      <td>0.973234</td>\n",
       "      <td>0.003339</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>33.490216</td>\n",
       "      <td>4.567255</td>\n",
       "      <td>0.270678</td>\n",
       "      <td>0.033435</td>\n",
       "      <td>None</td>\n",
       "      <td>300</td>\n",
       "      <td>{'max_depth': None, 'n_estimators': 300}</td>\n",
       "      <td>0.977558</td>\n",
       "      <td>0.975763</td>\n",
       "      <td>0.973944</td>\n",
       "      <td>0.966757</td>\n",
       "      <td>0.971249</td>\n",
       "      <td>0.973054</td>\n",
       "      <td>0.003777</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>3.185402</td>\n",
       "      <td>0.094999</td>\n",
       "      <td>0.218539</td>\n",
       "      <td>0.010736</td>\n",
       "      <td>90</td>\n",
       "      <td>10</td>\n",
       "      <td>{'max_depth': 90, 'n_estimators': 10}</td>\n",
       "      <td>0.980251</td>\n",
       "      <td>0.972172</td>\n",
       "      <td>0.976640</td>\n",
       "      <td>0.970350</td>\n",
       "      <td>0.964960</td>\n",
       "      <td>0.972875</td>\n",
       "      <td>0.005256</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>40.863298</td>\n",
       "      <td>1.409541</td>\n",
       "      <td>0.401632</td>\n",
       "      <td>0.077690</td>\n",
       "      <td>90</td>\n",
       "      <td>300</td>\n",
       "      <td>{'max_depth': 90, 'n_estimators': 300}</td>\n",
       "      <td>0.977558</td>\n",
       "      <td>0.974865</td>\n",
       "      <td>0.973046</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.970350</td>\n",
       "      <td>0.972875</td>\n",
       "      <td>0.003193</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>24.617707</td>\n",
       "      <td>0.246948</td>\n",
       "      <td>0.422405</td>\n",
       "      <td>0.026908</td>\n",
       "      <td>None</td>\n",
       "      <td>150</td>\n",
       "      <td>{'max_depth': None, 'n_estimators': 150}</td>\n",
       "      <td>0.976661</td>\n",
       "      <td>0.973070</td>\n",
       "      <td>0.973944</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.971249</td>\n",
       "      <td>0.972695</td>\n",
       "      <td>0.002708</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "7       21.955820      0.180804         0.339013        0.007180   \n",
       "11      33.490216      4.567255         0.270678        0.033435   \n",
       "6        3.185402      0.094999         0.218539        0.010736   \n",
       "8       40.863298      1.409541         0.401632        0.077690   \n",
       "10      24.617707      0.246948         0.422405        0.026908   \n",
       "\n",
       "   param_max_depth param_n_estimators  \\\n",
       "7               90                150   \n",
       "11            None                300   \n",
       "6               90                 10   \n",
       "8               90                300   \n",
       "10            None                150   \n",
       "\n",
       "                                      params  split0_test_score  \\\n",
       "7     {'max_depth': 90, 'n_estimators': 150}           0.975763   \n",
       "11  {'max_depth': None, 'n_estimators': 300}           0.977558   \n",
       "6      {'max_depth': 90, 'n_estimators': 10}           0.980251   \n",
       "8     {'max_depth': 90, 'n_estimators': 300}           0.977558   \n",
       "10  {'max_depth': None, 'n_estimators': 150}           0.976661   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "7            0.975763           0.973944           0.966757   \n",
       "11           0.975763           0.973944           0.966757   \n",
       "6            0.972172           0.976640           0.970350   \n",
       "8            0.974865           0.973046           0.968553   \n",
       "10           0.973070           0.973944           0.968553   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \n",
       "7            0.973944         0.973234        0.003339                1  \n",
       "11           0.971249         0.973054        0.003777                2  \n",
       "6            0.964960         0.972875        0.005256                3  \n",
       "8            0.970350         0.972875        0.003193                3  \n",
       "10           0.971249         0.972695        0.002708                5  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestClassifier()\n",
    "param = {'n_estimators': [10, 150, 300],\n",
    "        'max_depth': [30, 60, 90, None]}\n",
    "gs = GridSearchCV(rf, param, cv=5, n_jobs=-1)\n",
    "gs_fit = gs.fit(X_count_features, data['label'])\n",
    "pd.DataFrame(gs_fit.cv_results_).sort_values('mean_test_score', ascending=False)[0:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gradient Boosting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_learning_rate</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>238.344167</td>\n",
       "      <td>0.630772</td>\n",
       "      <td>0.318031</td>\n",
       "      <td>0.014571</td>\n",
       "      <td>0.1</td>\n",
       "      <td>7</td>\n",
       "      <td>150</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 7, 'n_esti...</td>\n",
       "      <td>0.965889</td>\n",
       "      <td>0.980251</td>\n",
       "      <td>0.970350</td>\n",
       "      <td>0.966757</td>\n",
       "      <td>0.969452</td>\n",
       "      <td>0.970540</td>\n",
       "      <td>0.005128</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>349.199861</td>\n",
       "      <td>4.021879</td>\n",
       "      <td>0.287572</td>\n",
       "      <td>0.056221</td>\n",
       "      <td>0.1</td>\n",
       "      <td>11</td>\n",
       "      <td>150</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 11, 'n_est...</td>\n",
       "      <td>0.964991</td>\n",
       "      <td>0.976661</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.969462</td>\n",
       "      <td>0.003855</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>287.632081</td>\n",
       "      <td>20.763992</td>\n",
       "      <td>0.300318</td>\n",
       "      <td>0.053156</td>\n",
       "      <td>0.1</td>\n",
       "      <td>15</td>\n",
       "      <td>100</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 15, 'n_est...</td>\n",
       "      <td>0.964991</td>\n",
       "      <td>0.974865</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.967655</td>\n",
       "      <td>0.969452</td>\n",
       "      <td>0.969103</td>\n",
       "      <td>0.003244</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>282.562305</td>\n",
       "      <td>35.462995</td>\n",
       "      <td>0.188738</td>\n",
       "      <td>0.009641</td>\n",
       "      <td>0.1</td>\n",
       "      <td>15</td>\n",
       "      <td>150</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 15, 'n_est...</td>\n",
       "      <td>0.964991</td>\n",
       "      <td>0.974865</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.966757</td>\n",
       "      <td>0.967655</td>\n",
       "      <td>0.968564</td>\n",
       "      <td>0.003364</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>227.757999</td>\n",
       "      <td>2.819332</td>\n",
       "      <td>0.317613</td>\n",
       "      <td>0.025518</td>\n",
       "      <td>0.1</td>\n",
       "      <td>11</td>\n",
       "      <td>100</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 11, 'n_est...</td>\n",
       "      <td>0.964093</td>\n",
       "      <td>0.975763</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.965858</td>\n",
       "      <td>0.964960</td>\n",
       "      <td>0.967845</td>\n",
       "      <td>0.004232</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "1     238.344167      0.630772         0.318031        0.014571   \n",
       "3     349.199861      4.021879         0.287572        0.056221   \n",
       "4     287.632081     20.763992         0.300318        0.053156   \n",
       "5     282.562305     35.462995         0.188738        0.009641   \n",
       "2     227.757999      2.819332         0.317613        0.025518   \n",
       "\n",
       "  param_learning_rate param_max_depth param_n_estimators  \\\n",
       "1                 0.1               7                150   \n",
       "3                 0.1              11                150   \n",
       "4                 0.1              15                100   \n",
       "5                 0.1              15                150   \n",
       "2                 0.1              11                100   \n",
       "\n",
       "                                              params  split0_test_score  \\\n",
       "1  {'learning_rate': 0.1, 'max_depth': 7, 'n_esti...           0.965889   \n",
       "3  {'learning_rate': 0.1, 'max_depth': 11, 'n_est...           0.964991   \n",
       "4  {'learning_rate': 0.1, 'max_depth': 15, 'n_est...           0.964991   \n",
       "5  {'learning_rate': 0.1, 'max_depth': 15, 'n_est...           0.964991   \n",
       "2  {'learning_rate': 0.1, 'max_depth': 11, 'n_est...           0.964093   \n",
       "\n",
       "   split1_test_score  split2_test_score  split3_test_score  split4_test_score  \\\n",
       "1           0.980251           0.970350           0.966757           0.969452   \n",
       "3           0.976661           0.968553           0.968553           0.968553   \n",
       "4           0.974865           0.968553           0.967655           0.969452   \n",
       "5           0.974865           0.968553           0.966757           0.967655   \n",
       "2           0.975763           0.968553           0.965858           0.964960   \n",
       "\n",
       "   mean_test_score  std_test_score  rank_test_score  \n",
       "1         0.970540        0.005128                1  \n",
       "3         0.969462        0.003855                2  \n",
       "4         0.969103        0.003244                3  \n",
       "5         0.968564        0.003364                4  \n",
       "2         0.967845        0.004232                5  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gb = GradientBoostingClassifier()\n",
    "param = {\n",
    "    'n_estimators': [100, 150], \n",
    "    'max_depth': [7, 11, 15],\n",
    "    'learning_rate': [0.1]\n",
    "}\n",
    "\n",
    "clf = GridSearchCV(gb, param, cv=5, n_jobs=-1)\n",
    "cv_fit = clf.fit(X_features, data['label'])\n",
    "pd.DataFrame(cv_fit.cv_results_).sort_values('mean_test_score', ascending=False)[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_learning_rate</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>404.235646</td>\n",
       "      <td>4.423075</td>\n",
       "      <td>0.296290</td>\n",
       "      <td>0.035830</td>\n",
       "      <td>0.1</td>\n",
       "      <td>11</td>\n",
       "      <td>150</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 11, 'n_est...</td>\n",
       "      <td>0.963196</td>\n",
       "      <td>0.978456</td>\n",
       "      <td>0.970350</td>\n",
       "      <td>0.965858</td>\n",
       "      <td>0.970350</td>\n",
       "      <td>0.969642</td>\n",
       "      <td>0.005188</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>267.181450</td>\n",
       "      <td>32.307128</td>\n",
       "      <td>0.191021</td>\n",
       "      <td>0.020828</td>\n",
       "      <td>0.1</td>\n",
       "      <td>15</td>\n",
       "      <td>150</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 15, 'n_est...</td>\n",
       "      <td>0.964093</td>\n",
       "      <td>0.979354</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.963163</td>\n",
       "      <td>0.971249</td>\n",
       "      <td>0.969282</td>\n",
       "      <td>0.005834</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>224.771290</td>\n",
       "      <td>0.802219</td>\n",
       "      <td>0.385718</td>\n",
       "      <td>0.112161</td>\n",
       "      <td>0.1</td>\n",
       "      <td>7</td>\n",
       "      <td>150</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 7, 'n_esti...</td>\n",
       "      <td>0.964093</td>\n",
       "      <td>0.978456</td>\n",
       "      <td>0.969452</td>\n",
       "      <td>0.964061</td>\n",
       "      <td>0.965858</td>\n",
       "      <td>0.968384</td>\n",
       "      <td>0.005405</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>261.352368</td>\n",
       "      <td>36.483303</td>\n",
       "      <td>0.334987</td>\n",
       "      <td>0.023070</td>\n",
       "      <td>0.1</td>\n",
       "      <td>11</td>\n",
       "      <td>100</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 11, 'n_est...</td>\n",
       "      <td>0.963196</td>\n",
       "      <td>0.976661</td>\n",
       "      <td>0.965858</td>\n",
       "      <td>0.963163</td>\n",
       "      <td>0.968553</td>\n",
       "      <td>0.967486</td>\n",
       "      <td>0.005001</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>329.566580</td>\n",
       "      <td>50.065862</td>\n",
       "      <td>0.307354</td>\n",
       "      <td>0.051345</td>\n",
       "      <td>0.1</td>\n",
       "      <td>15</td>\n",
       "      <td>100</td>\n",
       "      <td>{'learning_rate': 0.1, 'max_depth': 15, 'n_est...</td>\n",
       "      <td>0.962298</td>\n",
       "      <td>0.973070</td>\n",
       "      <td>0.967655</td>\n",
       "      <td>0.962264</td>\n",
       "      <td>0.969452</td>\n",
       "      <td>0.966948</td>\n",
       "      <td>0.004191</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "3     404.235646      4.423075         0.296290        0.035830   \n",
       "5     267.181450     32.307128         0.191021        0.020828   \n",
       "1     224.771290      0.802219         0.385718        0.112161   \n",
       "2     261.352368     36.483303         0.334987        0.023070   \n",
       "4     329.566580     50.065862         0.307354        0.051345   \n",
       "\n",
       "  param_learning_rate param_max_depth param_n_estimators  \\\n",
       "3                 0.1              11                150   \n",
       "5                 0.1              15                150   \n",
       "1                 0.1               7                150   \n",
       "2                 0.1              11                100   \n",
       "4                 0.1              15                100   \n",
       "\n",
       "                                              params  split0_test_score  \\\n",
       "3  {'learning_rate': 0.1, 'max_depth': 11, 'n_est...           0.963196   \n",
       "5  {'learning_rate': 0.1, 'max_depth': 15, 'n_est...           0.964093   \n",
       "1  {'learning_rate': 0.1, 'max_depth': 7, 'n_esti...           0.964093   \n",
       "2  {'learning_rate': 0.1, 'max_depth': 11, 'n_est...           0.963196   \n",
       "4  {'learning_rate': 0.1, 'max_depth': 15, 'n_est...           0.962298   \n",
       "\n",
       "   split1_test_score  split2_test_score  split3_test_score  split4_test_score  \\\n",
       "3           0.978456           0.970350           0.965858           0.970350   \n",
       "5           0.979354           0.968553           0.963163           0.971249   \n",
       "1           0.978456           0.969452           0.964061           0.965858   \n",
       "2           0.976661           0.965858           0.963163           0.968553   \n",
       "4           0.973070           0.967655           0.962264           0.969452   \n",
       "\n",
       "   mean_test_score  std_test_score  rank_test_score  \n",
       "3         0.969642        0.005188                1  \n",
       "5         0.969282        0.005834                2  \n",
       "1         0.968384        0.005405                3  \n",
       "2         0.967486        0.005001                4  \n",
       "4         0.966948        0.004191                5  "
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gb = GradientBoostingClassifier()\n",
    "param = {\n",
    "    'n_estimators': [100, 150], \n",
    "    'max_depth': [7, 11, 15],\n",
    "    'learning_rate': [0.1]\n",
    "}\n",
    "\n",
    "clf = GridSearchCV(gb, param, cv=5, n_jobs=-1)\n",
    "cv_fit = clf.fit(X_count_features, data['label'])\n",
    "pd.DataFrame(cv_fit.cv_results_).sort_values('mean_test_score', ascending=False)[0:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model selection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Split data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_features, data['label'], test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_recall_fscore_support as score\n",
    "import time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0.08097781387894042, 'body_length'),\n",
       " (0.04264358493114639, 1803),\n",
       " (0.03579127459914742, 3134),\n",
       " (0.032309364041029943, 4796),\n",
       " (0.015459259498580695, 5988),\n",
       " (0.01500241244626518, 6285),\n",
       " (0.014950072233750672, 5724),\n",
       " (0.014651211377860334, 2171),\n",
       " (0.014145582159452049, 7461),\n",
       " (0.014133693141060057, 392)]"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestClassifier(n_estimators=10, max_depth=90, n_jobs=-1)\n",
    "\n",
    "start=time.time()\n",
    "rf_model = rf.fit(X_train, y_train)\n",
    "end=time.time()\n",
    "fit_time=(end-start)\n",
    "\n",
    "sorted(zip(rf_model.feature_importances_, X_train.columns), reverse=True)[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fit time: 4.375 / Predict time: 1.384 ---- Precision: 0.984 / Recall: 0.814 / Accuracy: 0.972\n"
     ]
    }
   ],
   "source": [
    "start = time.time()\n",
    "y_pred = rf_model.predict(X_test)\n",
    "end = time.time()\n",
    "pred_time = (end - start)\n",
    "\n",
    "precision, recall, fscore, train_support = score(y_test, y_pred, pos_label='spam', average='binary')\n",
    "print('Fit time: {} / Predict time: {} ---- Precision: {} / Recall: {} / Accuracy: {}'.format(\n",
    "    round(fit_time, 3), round(pred_time, 3), round(precision, 3), round(recall, 3), round((y_pred==y_test).sum()/len(y_pred), 3)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Gradient Boost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fit time: 781.057 / Predict time: 0.376 ---- Precision: 0.901 / Recall: 0.872 / Accuracy: 0.969\n"
     ]
    }
   ],
   "source": [
    "gb = GradientBoostingClassifier(n_estimators=150, max_depth=11)\n",
    "\n",
    "start = time.time()\n",
    "gb_model = gb.fit(X_train, y_train)\n",
    "end = time.time()\n",
    "fit_time = (end - start)\n",
    "\n",
    "start = time.time()\n",
    "y_pred = gb_model.predict(X_test)\n",
    "end = time.time()\n",
    "pred_time = (end - start)\n",
    "\n",
    "precision, recall, fscore, train_support = score(y_test, y_pred, pos_label='spam', average='binary')\n",
    "print('Fit time: {} / Predict time: {} ---- Precision: {} / Recall: {} / Accuracy: {}'.format(\n",
    "    round(fit_time, 3), round(pred_time, 3), round(precision, 3), round(recall, 3), round((y_pred==y_test).sum()/len(y_pred), 3)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "\n",
    "\n",
    "|   model/score  | Fit time | Predict time | Precision | Recall | Accuracy |\n",
    "|:--------------:|:--------:|:------------:|:---------:|:------:|:--------:|\n",
    "|  Random Forest |   4.375  |     1.384    |   0.984   |  0.814 |   0.972  |\n",
    "| Gradient Boost |  781.057 |     0.376    |   0.901   |  0.872 |   0.969  |\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-info\" style=\"padding:0px 10px; border-radius:5px;\"><h3 style='margin:10px 5px'> Inferences:</h3>\n",
    "</div>\n",
    "\n",
    "We can see final model perfomance in the table above. Fitting the data for Gradient Boost is taking  much longer time in comparison with Random Forest model, however the Predict Time is what we focused on. And you see, clearly, that Gradient Boost Model is making prediction spending 4 times fewer time than Random Forest.\n",
    "The target of this project is to detect spam text messeges,  therefore, we should seeking for better Recall score, than Precision, not speaking about Accuracy here. \n",
    "So, it turns out Random Forest better detects text as spam if it actually was a spam, but Gradient Boost is better on correctly identifying spam if it was labeled as spam. \n",
    "That saying, Gradient Boost model is incorrectly  detecting 12.8% of spam, meanwhile Random Forest is 18.6%."
   ]
  }
 ],
 "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}