Assingment 2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ML Summer School - Assignment 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Welcome to Machine Learning Summer School (2018) - Assignment 2. In this, assignment you will be making a spam classifer. That is given a SMS, you have to build a model that classifies whether the message is a spam or ham (not spam). "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Go through this Notebook very carefully. There are links to resources given in between, it is expected you go through them."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first step we will do is to import the machine learning libraries. We will be using NumPy, Matplotlib, Pandas, and scikit-learn in this assignment. We will also be using NLTK (a library widely used for text processing)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Importing libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from nltk.corpus import stopwords\n",
    "import pandas as pd\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import f1_score\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "import warnings\n",
    "warnings.simplefilter(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loading data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TASK 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first task you need to do is load the csv file as a Pandas DataFrame. First go ahead and open `sms-spam.csv` file in a text editor. Analyse the data - how many columns are there, how big the data is. Identify some patterns which may help the algorithm in classififcation.   \n",
    "\n",
    "Load the csv as a Pandas DataFrame. The Pandas DataFrame should be named `df` and must contain only two columns named `text` and `is spam`. The `text` column should contain the SMS and `is spam` column should contain `1` if the SMS is spam and `0` if the SMS is a ham.\n",
    "\n",
    "\n",
    "You will need to use Pandas' `read_csv` function. The csv file is in `latin1` encoding. You will need to care take of that. You may find `encoding`, `usecols` and `names` arguments of `read_csv` helpful.  \n",
    "To convert 'spam' and 'ham' to 1 and 0 respectively, you may find `DataFrame.apply` useful."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>is spam</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   is spam                                               text\n",
       "1        0  Go until jurong point, crazy.. Available only ...\n",
       "2        0                      Ok lar... Joking wif u oni...\n",
       "3        1  Free entry in 2 a wkly comp to win FA Cup fina...\n",
       "4        0  U dun say so early hor... U c already then say...\n",
       "5        0  Nah I don't think he goes to usf, he lives aro..."
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# YOUR CODE STARTS HERE\n",
    "\n",
    "df=pd.read_csv('sms-spam.csv',sep=',',usecols=[0,1],names=['is spam','text'],encoding=\"latin1\")\n",
    "df=df.drop(df.index[[0]])\n",
    "def classify(name):\n",
    "    if name==\"spam\":\n",
    "        return 1\n",
    "    else:\n",
    "        return 0\n",
    "df['is spam']=df['is spam'].apply(classify)\n",
    "# YOUR CODE ENDS HERE\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analysing data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The second step in Machine Learning Pipeline is to analyse the data. Analysing the data helps in [Feature Engineering](https://medium.com/mindorks/what-is-feature-engineering-for-machine-learning-d8ba3158d97a).  \n",
    "\n",
    "The most important part of traditional Machine Learning is how well you represent the data. Tradional Machine Learning approaches can't work directly on raw data. You need to represent the data in a way that is easy for machines to work with.\n",
    "\n",
    "In this section, we will analyse the data and side by side engineer new features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 5572 entries, 1 to 5572\n",
      "Data columns (total 2 columns):\n",
      "is spam    5572 non-null int64\n",
      "text       5572 non-null object\n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 130.6+ KB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that we don't have any data misising. So, we don't need to [handle missing data](https://machinelearningmastery.com/handle-missing-data-python/)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"4\" halign=\"left\">text</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>unique</th>\n",
       "      <th>top</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>is spam</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4825</td>\n",
       "      <td>4516</td>\n",
       "      <td>Sorry, I'll call later</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>747</td>\n",
       "      <td>653</td>\n",
       "      <td>Please call our customer service representativ...</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         text                                                               \n",
       "        count unique                                                top freq\n",
       "is spam                                                                     \n",
       "0        4825   4516                             Sorry, I'll call later   30\n",
       "1         747    653  Please call our customer service representativ...    4"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('is spam').describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[`groupby`](https://www.datacamp.com/community/tutorials/pandas-split-apply-combine-groupby) is a powerful way to analyse (and manipulate the data).  \n",
    "Using the above command, we can see that 747 messages are spam, while 4825 messages are ham. We also observe that most of the message are unique."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TASK 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot a pie chart showing the number of spam and ham messages in the dataset. Make sure you use you give it a title and use change the default colors (green or red maybe?)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3dc797f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# YOUR CODE STARTS HERE\n",
    "hamcount=df.groupby('is spam').describe().iat[0,0]\n",
    "spamcount=df.groupby('is spam').describe().iat[1,0]\n",
    "plt.pie([hamcount,spamcount],labels=['HAM','SPAM'],colors=['red','blue'],autopct='%1.1f%%',shadow=True,startangle=140,explode=[0,0.1])\n",
    "plt.title(\"SPAM vs HAM\")\n",
    "plt.show()\n",
    "# YOUR CODE ENDS HERE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Adding features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Percentage digits"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you have gone through the dataset, you would have observed that spam messages have a lot more digits in them as compared to ham messages. So, lets create a feature based on that."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>is spam</th>\n",
       "      <th>text</th>\n",
       "      <th>percentage digits</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>16.129032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   is spam                                               text  \\\n",
       "1        0  Go until jurong point, crazy.. Available only ...   \n",
       "2        0                      Ok lar... Joking wif u oni...   \n",
       "3        1  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "4        0  U dun say so early hor... U c already then say...   \n",
       "5        0  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "   percentage digits  \n",
       "1           0.000000  \n",
       "2           0.000000  \n",
       "3          16.129032  \n",
       "4           0.000000  \n",
       "5           0.000000  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def percentage_digits(nnn):\n",
    "    counter = 0\n",
    "    for i in nnn:\n",
    "        if i.isdigit():\n",
    "            counter += 1\n",
    "    return (counter / len(nnn)) * 100\n",
    "\n",
    "df['percentage digits'] = df['text'].apply(percentage_digits)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's plot to check if our observation was correct."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3dc797240>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 0]['percentage digits'], bins=10, range=(0, 20), rwidth=0.8)\n",
    "plt.xlabel('Percentage digits')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage digits of non-spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3dc884898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 1]['percentage digits'], bins=10, range=(0, 20), rwidth=0.8)\n",
    "plt.xlabel('Percentage digits')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage digits of spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Indeed it was correct. Therefore, we can use this feature for our machine learning model. Let us create a bunch of more features like this."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Percentage question mark"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>is spam</th>\n",
       "      <th>text</th>\n",
       "      <th>percentage digits</th>\n",
       "      <th>percentage question mark</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>16.129032</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   is spam                                               text  \\\n",
       "1        0  Go until jurong point, crazy.. Available only ...   \n",
       "2        0                      Ok lar... Joking wif u oni...   \n",
       "3        1  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "4        0  U dun say so early hor... U c already then say...   \n",
       "5        0  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "   percentage digits  percentage question mark  \n",
       "1           0.000000                       0.0  \n",
       "2           0.000000                       0.0  \n",
       "3          16.129032                       0.0  \n",
       "4           0.000000                       0.0  \n",
       "5           0.000000                       0.0  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def percentage_question_mark(nnn):\n",
    "    counter = 0\n",
    "    for i in nnn:\n",
    "        if i == '?':\n",
    "            counter += 1\n",
    "    return (counter / len(nnn)) * 100\n",
    "\n",
    "df['percentage question mark'] = df['text'].apply(percentage_question_mark)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3ffecf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 1]['percentage question mark'], bins=10, range=(0, 4), rwidth=0.8)\n",
    "plt.xlabel('Percentage question mark')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage question mark of spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaoAAAEWCAYAAAA3h9P4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xm8HFWZ//HPlxD2JWACZoOwRBCUzbD4U5FFWRVQccQRCYhGBERmGAWUkU2U0VEWlSUKAqIsImJAEBEIyCgkAcISAhIDSEggYQuE1YTn98c5TSqd7nsrye3uSu73/Xr161ad2p6qrqqn69S5VYoIzMzMqmq5TgdgZmbWFScqMzOrNCcqMzOrNCcqMzOrNCcqMzOrNCcqMzOrtI4nKknnSfrvHprXepLmSOqT+8dK+mJPzDvP7wZJI3tqfouw3O9IelbS0+1etvUcSZMk7dTpOBaVpJMkXdpD8/K+bIts+VbOXNLjwLrAXGAe8BBwCTA6It4CiIjDFmFeX4yIPzcbJyL+Cay2ZFG/vbyTgI0j4sDC/PfsiXkvYhxDgWOA9SNiZruX31MkjQUujYifdzqWdpB0ETAtIk6olUXE5p2LqPOWlX3Z2q8dV1Qfj4jVgfWB04FjgQt6eiGSWpp0O2h94Ll2HNjL8Da0xdTD+0Tb9mVbxkREyz7A48BH6sq2A94C3pP7LwK+k7v7A9cBLwLPA38hJdNf5mleA+YA3wCGAQEcCvwTuL1Qtnye31jge8A4YDbwe2DtPGwn0i/eheIF9gDeBP6Vl3dfYX5fzN3LAScATwAzSVeKa+ZhtThG5tieBb7VxXZaM08/K8/vhDz/j+R1fivHcVGDaXcCpgHfzMt5HPhcYfiKwP/mOJ4BzgNWrpv2WOBp4Je5fF9gIvAS8A9gj0KcFwAzgKeA7wB98rCDgTvysl4AHgP2zMNOI11Rv57X4ye5/Czgybycu4EPFeJeGbg4z2ty/s6nFYYPAn6bt9ljwFFdbN93AGPycsYBpwJ31H1XyxfGf/t7zv1fyDG8ANxIuiIAEHBG/v5nA/cD7wFGkfadN/P6Xlt/POTv5Uxgev6cCaxY970ck+c9Aziki/Ubm7+Lv9aWl9f5V3mdxwPDCuN3td1PAq4CLs3Dv5jLLs3D+wKX5W2/Qov25Ybr3Wze3e1/TbbZxsBt+Xt7FriiMCyAo4CpedgPCsvZCLgFeC4P+xXQr+4c8vW8L7xCOl7WBW4AXgb+DKzVJKba+n+jsP77AXsBfyedE79ZGH854DjSMfoccCXzz28r5e/wOdL5dDywbmFbTc3xPEY+X5RYt22Ae/N0vwGuIJ+78/CPkc4bL5L2xS0Kw44lnTNeBh4Bdl2kXNITCamLneFx6hJVLv8n8JXcfRHzE9X3SCfSvvnzIUCN5sX8E8wlwKqkE1utrJioniKdPFYlHVy1A24nmiSqwgF7aYMTQi1RfQGYAmxIqm68mvkn+locP8txbQm8Aby7yXa6hJREV8/T/h04tFmcDXbuucCPSCe/D5MOkE3y8DNJJ+m18/yvBb5XN+3/5GlXJv2QmA18lHQgDAY2zeNfA5yft+U6pJP+lws7/7+ALwF9gK+QTsCq33aF2A8knVCXJ52cngZWysNOJ51I1gKGkA78aYUD9G7g28AK+TuYCuzeZBtdTjqIV837wlOUTFSkE8UU4N05zhOAv+Zhu+c4+pGS1ruBgfX7dZP96xTgzrwdB5AO7FPrvpdTSMfBXsCrND/Bjc0xbkQ6mT9E2oc+kmO+BPhFye1+Uv4e98vbeeVcdmnu/kNetz4t3Jcbrnc38z6YLva/Bsu6DPhWXseVgA8WhgVwK+mYWS8vp7Y/bEw6NlbM39vtwJl13/GdpOQ0mJRw7gG2ztPcApzYzfp/O6//l0hJ+dd5nTcn/djbMI9/dF7WkDzv84HL8rAvk471VfL2eB+wBukYeIn554eBwObdrRvpOHsC+FqO7ZOkH2K1c/c2eV23z8sbmbfFisAmpB9GgwrH3EaLlEsWJwGVnnnzRHUn+QqDBRPVKXlH3Li7eTH/BLNhg7Jiojq9MHyzvHH7sOSJ6mbg8MKwTUgHyvKFOIYUho8DDmiwXn1ISWyzQtmXgbGLeHCvWii7Evhv0snzleJOAbwfeKww7Zvkk1QuOx84o8Fy1s1xrlwo+yxwa+FEMaUwbJW8Dd5Zv+26WJcXgC1z9wKJh/TLvpaotgf+WTft8RROxnXb91/kZJvLvkv5RHUD+WSY+5cjnTzXB3YhncR2IP/iLox3EV0nqn8AexWG7Q48XvheXquLaSawQ5PtNpbCFTvwQ+CGQv/HgYklt/tJwO11w08i/di5DTib5if/ntiXG653iXl3uf81WNYlwGgKx2hhWJBrEXL/4cDNTeazH3Bv3XdcrNH4LXBuof+rwDXdrH+tlmL1HMv2hXHuBvbL3ZMpXJmQkk7tHPQF6q5q8jirkq54PkXhWO5u3YAdST/wVBh+B/PP3eeSf2gVhj9C+uG8cf4ePwL07WqZzT6davU3mHQZW+8HpF+Gf5I0VdJxJeb15CIMf4L0a6B/qSi7NijPrzjv5Ukn9Jpiy6ZXadzQoz/zf60U5zV4EWJ5ISJeqZt+EOlX0SrA3ZJelPQi8MdcXjMrIl4v9A8lnUTrrU/adjMK8zqfdEVQ8/b6RsSrubNp4xZJx0iaLGl2nt+azP9uBrHgd1fsXh8YVIsjT/tNFtz2NQNI30v9flDW+sBZheU8T/oBMDgibgF+AvwUeEbSaElrlJxvo/1nUKH/uYiYW+hvtv/UPFPofq1B/9vTdrPdofExtQOwBemHXzSJoSf25WbrXWbeDfc/SR/KrYHnSJqUy79B+h7H5daYX6iLo35/GQQgaR1Jl0t6StJLpCvN+vNJ6e+igeciYl5h3Ebzq02/PvC7wr45mVTFvi7pdsmNwOWSpkv6vqS++TzxGeAw0rH8B0mblli3QcBTdd99/TF5TN0xOZR0FTWFdPV3EjAzL6O4r3er7YlK0rakneuO+mER8XJEHBMRG5J+Bf6npF1rg5vMsll5zdBC93qkXxzPkq40VinE1YcFT+DdzXc66cspznsuC+5UZTybY6qf11OLMI+1JK1aN/30PO/XSJf2/fJnzYgoHij16/kkqQqp3pOkX7T9C/NaI8q3ZFtgOZI+RKq3/jdS1U4/UpWj8igzSFUaNcXv8UnSVWG/wmf1iNirwXJnkb6X+v2gppbgVymUvbNuWV+uW9bKEfFXgIg4OyLeR6qWeRfp/sRC69tAo/1nejfTLLES2x0ax/4nUtX8zZIa/SCAntmXm1nseUfEXyJitfzZPJc9HRFfiohBpCuzcyRtXJisfn+pfTffI22fLSJiDVI1anHbtdOTpPtwxX1zpYh4KiL+FREnR8RmwP8j3T86CCAiboyIj5KuwB4m3aKArtdtBjBYUnFd64/J0+piWSUiLsvL/HVEfJD0/QXpdkNpbUtUktaQ9DHS/YJLI+KBBuN8TNLGeWO8RPp1UPt18QzpXsSiOlDSZpJWIVUtXpV/sfwdWEnS3pL6ku49rFiY7hlgmKRm2+gy4D8kbSBpNVJ10hV1vwa7lWO5EjhN0uqS1gf+k/RrZlGcLGmFfCL6GPCbSP8C8DPgDEnrAEgaLGn3LuZzAXCIpF0lLZfH3zQiZpBOVj/M3+VykjaS9OGS8dV/f6uTEsgsYHlJ3ybVoddcCRwvaS1Jg4EjC8PGAS9JOlbSypL6SHpP/hG0gLx9rwZOkrSKpM1I9ee14bNIJ7sD83y+wIKJ+rwcx+YAktaU9Oncva2k7fP+8wrp/kHZ/fUy4ARJAyT1J92X6JH/VepGd9u9qYj4Pul+yc055vrhPbUvN1p2j85b0qcl1X4IvUA6ec4rjPL1vO8NJd2XuSKXr05qDPJi3i+/TuecR9oe6wPkfWnf3L2zpPfmH+AvkZL8PEnrSton/7B9g7QutfXuat3+lsc7UtLyeTnbFYb/DDgsHw+StGo+t64uaRNJu0hakXSMvMaC27pb7UhU10p6mZRxv0W66X9Ik3GHk1rFzCFtmHMiYmwe9j3Sgf2ipP9ahOX/knS/4GnSTdOjACJiNqnu+eekE9UrpBY3Nb/Jf5+TdE+D+V6Y5307qeXM66T658Xx1bz8qaQrzV/n+Zf1NOlgm05qqXNYRDychx1Lqk69M1/O/5l0P62hiBhH+n7OIP3Svo35v2IPIlW/PJSXdxXpV1kZZwH7S3pB0tmkaokbSD8YniBtv2JVwimk7+OxHPNVpAOrdtL6OLBVHv4s6Xtcs8myjyRVlzxN2hd+UTf8S6SD8jnSldFfawMi4nekX3+X5+33IFD7f7o1SAfoC3kdniO1OoOU8DfL++s1DWL6DjCB1EjkAdIN9+80ib8ndbfduxQRp5Ia1fxZ0toNRlnSfbkrPTnvbYG7JM0h3X/7WkQ8Vhj+e9L9oImkBiS1f6k5mdRwYHYuv3oxl98TziLF/qd8jr2TdP8WUq3AVaQkNZl0HF9KOucfQzpXPE+6h3R4nqbpukXEm6QGFIeS7nEdSGqhXTsmJ5COo5+QjocppPuGkC4ATicdp0+Tbhd8E0DS5zS/OrapWossW0opPeng0ogY0t24SzNJXyE1Ril7BdfVvA4mNZb44BIHZsscSQEMz/dWrAlJdwHnRUT9D78e1/FHKJk1ImmgpA/kKsZNSL8Cf9fpuMx6K0kflvTOXPU3ktS45o/tWLafRGBVtQKpVeEGpKqGy4FzOhqRWe+2Cek+4WqklsH753vXLeeqPzMzqzRX/ZmZWaUtk1V//fv3j2HDhnU6DDOzpcrdd9/9bEQM6H7M9lomE9WwYcOYMGFCp8MwM1uqSFqUp7a0jav+zMys0pyozMys0pyozMys0pyozMys0pyozMys0pyozMys0pyozMys0pyozMys0pyozMys0pbJJ1MsqWHH/aHly3j89L1bvgwzs2WBr6jMzKzSnKjMzKzSnKjMzKzSnKjMzKzSnKjMzKzSnKjMzKzSnKjMzKzSnKjMzKzSnKjMzKzSWpaoJK0kaZyk+yRNknRyLr9I0mOSJubPVrlcks6WNEXS/ZK2KcxrpKRH82dkq2I2M7PqaeUjlN4AdomIOZL6AndIuiEP+3pEXFU3/p7A8PzZHjgX2F7S2sCJwAgggLsljYmIF1oYu5mZVUTLrqgimZN7++ZPdDHJvsAlebo7gX6SBgK7AzdFxPM5Od0E7NGquM3MrFpaeo9KUh9JE4GZpGRzVx50Wq7eO0PSirlsMPBkYfJpuaxZuZmZ9QItTVQRMS8itgKGANtJeg9wPLApsC2wNnBsHl2NZtFF+QIkjZI0QdKEWbNm9Uj8ZmbWeW1p9RcRLwJjgT0iYkau3nsD+AWwXR5tGjC0MNkQYHoX5fXLGB0RIyJixIABA1qwFmZm1gmtbPU3QFK/3L0y8BHg4XzfCUkC9gMezJOMAQ7Krf92AGZHxAzgRmA3SWtJWgvYLZeZmVkv0MpWfwOBiyX1ISXEKyPiOkm3SBpAqtKbCByWx78e2AuYArwKHAIQEc9LOhUYn8c7JSKeb2HcZmZWIS1LVBFxP7B1g/JdmowfwBFNhl0IXNijAZqZ2VLBT6YwM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKc6IyM7NKa1mikrSSpHGS7pM0SdLJuXwDSXdJelTSFZJWyOUr5v4pefiwwryOz+WPSNq9VTGbmVn1tPKK6g1gl4jYEtgK2EPSDsD/AGdExHDgBeDQPP6hwAsRsTFwRh4PSZsBBwCbA3sA50jq08K4zcysQlqWqCKZk3v75k8AuwBX5fKLgf1y9765nzx8V0nK5ZdHxBsR8RgwBdiuVXGbmVm1tPQelaQ+kiYCM4GbgH8AL0bE3DzKNGBw7h4MPAmQh88G3lEsbzBNcVmjJE2QNGHWrFmtWB0zM+uAliaqiJgXEVsBQ0hXQe9uNFr+qybDmpXXL2t0RIyIiBEDBgxY3JDNzKxi2tLqLyJeBMYCOwD9JC2fBw0BpufuacBQgDx8TeD5YnmDaczMbBnXylZ/AyT1y90rAx8BJgO3Avvn0UYCv8/dY3I/efgtERG5/IDcKnADYDgwrlVxm5lZtSzf/SiLbSBwcW6htxxwZURcJ+kh4HJJ3wHuBS7I418A/FLSFNKV1AEAETFJ0pXAQ8Bc4IiImNfCuM3MrEJalqgi4n5g6wblU2nQai8iXgc+3WRepwGn9XSMZmZWfd1W/UnaSNKKuXsnSUfVqvTMzMxarcw9qt8C8yRtTKqe2wD4dUujMjMzy8okqrfy/zV9AjgzIv6DdP/JzMys5cokqn9J+iypRd51uaxv60IyMzObr0yiOgR4P3BaRDyWm4hf2tqwzMzMkm5b/UXEQ5KOBdbL/Y8Bp7c6MDMzMyjX6u/jwETgj7l/K0ljWh2YmZkZlKv6O4n0f08vAkTERFLLPzMzs5Yrk6jmRsTsurKFHgprZmbWCmWeTPGgpH8H+kgaDhwF/LW1YZmZmSVlrqi+Snq77hvAZcBLwNGtDMrMzKymTKu/V4Fv5Y+ZmVlbdZuoJF3LwvekZgMTgPPzw2TNzMxaokzV31RgDvCz/HkJeAZ4V+43MzNrmTKNKbaOiB0L/ddKuj0idpQ0qVWBmZmZQbkrqgGS1qv15O7+uffNlkRlZmaWlbmiOga4Q9I/AJH+2fdwSasCF7cyODMzszKt/q7P/z+1KSlRPVxoQHFmK4MzMzMr+yr64cAmwErAFpKIiEtaF5aZmVlS5qG0JwI/zp+dge8D+5SYbqikWyVNljRJ0tdy+UmSnpI0MX/2KkxzvKQpkh6RtHuhfI9cNkXScYuxnmZmtpQqc0W1P7AlcG9EHCJpXeDnJaabCxwTEfdIWh24W9JNedgZEfG/xZElbQYcQHoKxiDgz5LelQf/FPgoMA0YL2lMRDxUIgYzM1vKlUlUr0XEW5LmSloDmAls2N1EETEDmJG7X5Y0GRjcxST7ApdHxBvAY5KmkJ7aDjAlIqYCSLo8j+tEZWbWC5Rpnj5BUj/SP/feDdwDjFuUhUgaBmwN3JWLjpR0v6QLJa2VywYDTxYmm5bLmpXXL2OUpAmSJsyaNWtRwjMzswrrNlFFxOER8WJEnEeqfhsZEYeUXYCk1YDfAkdHxEvAucBGwFakK64f1kZttPguyuvjHB0RIyJixIABA8qGZ2ZmFVemMcUH8v9MAXwQOFjS+mVmLqkvKUn9KiKuBoiIZyJiXkS8RbpKq1XvTQOGFiYfAkzvotzMzHqBMlV/5wKvStoS+AbwBNBt03RJAi4AJkfEjwrlAwujfQJ4MHePAQ6QtKKkDUhN4scB44HhkjaQtAKpwcWYEnGbmdkyoExjirkREZL2Bc6KiAskjSwx3QeAzwMPSJqYy74JfFbSVqTqu8eBLwNExCRJV5IaScwFjoiIeQCSjgRuBPoAF0aEnzFoZtZLlElUL0s6HjgQ2FFSH6BvdxNFxB00vr90fRfTnAac1qD8+q6mMzOzZVeZqr/PkN7ue2hEPE1qcfeDlkZlZmaWlbqiIlX5zcv/gLsp6ZX0ZmZmLVfmiup2YEVJg4GbgUOAi1oZlJmZWU2ZRKWIeBX4JPDjiPgE6TFHZmZmLVcqUUl6P/A54A+5rE/rQjIzM5uvTKI6Gjge+F1uQr4hcGtrwzIzM0vKvDjxNuC22tMp8sNhj2p1YGZmZlDuEUrvl/QQMDn3bynpnJZHZmZmRrmqvzOB3YHnACLiPmDHVgZlZmZWUyZRERFP1hXNa0EsZmZmCynzD79PSvp/QOSHwh5FrgY0MzNrtTJXVIcBR5AenTSN9B6pI1oZlJmZWU2ZVn/Pkv6HyszMrO26TVT53VBfBYYVx4+IfVoXlpmZWVLmHtU1pBcgXgu81dpwzMzMFlQmUb0eEWe3PBIzM7MGyiSqsySdCPyJ9F4qACLinpZFZWZmlpVJVO8lvVJ+F+ZX/UXuNzMza6kyzdM/AWwYER+OiJ3zp9skJWmopFslTZY0SdLXcvnakm6S9Gj+u1Yul6SzJU2RdL+kbQrzGpnHf1TSyMVdWTMzW/qUSVT3Af0WY95zgWMi4t3ADsARkjYDjgNujojhpBcxHpfH3xMYnj+jgHMhJTbgRGB7YDvgxFpyMzOzZV+Zqr91gYcljWfBe1RdNk+PiBnAjNz9sqTJpH8a3hfYKY92MTAWODaXXxIRAdwpqZ+kgXncmyLieQBJNwF7AJeVW0UzM1ualUlUJy7pQiQNA7YG7gLWzUmMiJghaZ082mCg+EzBabmsWXn9MkaRrsRYb731ljRkMzOriLLvo1psklYDfgscHREvSWo6aqPFd1G+YEHEaGA0wIgRIxYabmZmS6dST09fXJL6kpLUryLi6lz8TK7SI/+dmcunAUMLkw8BpndRbmZmvUDLEpXSpdMFwOSI+FFh0Big1nJvJPD7QvlBufXfDsDsXEV4I7CbpLVyI4rdcpmZmfUCTROVpJvz3/9ZzHl/gPz/V5Im5s9ewOnARyU9Cnw09wNcD0wFpgA/Aw4HyI0oTgXG588ptYYVZma27OvqHtVASR8G9pF0OXX3irp7MkVE3FE/TcGuDcYPmrw+JCIuBC7sanlmZrZs6ipRfZv0P05DgB/VDfOTKczMrC2aJqqIuAq4StJ/R8SpbYzJzMzsbWWap58qaR9gx1w0NiKua21YZmZmSbet/iR9D/ga8FD+fC2XmZmZtVyZJ1PsDWwVEW8BSLoYuBc4vpWBmZmZQfn/oyo+lHbNVgRiZmbWSJkrqu8B90q6ldTcfEd8NWVmZm1SpjHFZZLGAtuSEtWxEfF0qwMzMzODcldUtVd2jGlxLGZmZgtp6UNpzczMlpQTlZmZVVqXiUrScpIebFcwZmZm9bpMVPl/p+6T5FfmmplZR5RpTDEQmCRpHPBKrTAi9mlZVGZmZlmZRHVyy6MwMzNrosz/Ud0maX1geET8WdIqQJ/Wh2ZmZlbuobRfAq4Czs9Fg4FrWhmUmZlZTZnm6UeQXiv/EkBEPAqs08qgzMzMasokqjci4s1aj6TlSW/4NTMza7kyieo2Sd8EVpb0UeA3wLXdTSTpQkkzi/+HJekkSU9Jmpg/exWGHS9piqRHJO1eKN8jl02RdNyirZ6ZmS3tyiSq44BZwAPAl4HrgRNKTHcRsEeD8jMiYqv8uR5A0mbAAcDmeZpzJPWR1Af4KbAnsBnw2TyumZn1EmVa/b2VX5Z4F6nK75GI6LbqLyJulzSsZBz7ApdHxBvAY5KmANvlYVMiYiqApMvzuA+VnK+ZmS3lyrT62xv4B3A28BNgiqQ9l2CZR0q6P1cNrpXLBgNPFsaZlsualTeKc5SkCZImzJo1awnCMzOzKilT9fdDYOeI2CkiPgzsDJyxmMs7F9gI2AqYkecN6T1X9aKL8oULI0ZHxIiIGDFgwIDFDM/MzKqmzJMpZkbElEL/VGDm4iwsIp6pdUv6GXBd7p0GDC2MOgSYnrublZuZWS/QNFFJ+mTunCTpeuBK0tXMp4Hxi7MwSQPzSxgBPgHUWgSOAX4t6UfAIGA4MI50RTVc0gbAU6QGF/++OMs2M7OlU1dXVB8vdD8DfDh3zwLWWnj0BUm6DNgJ6C9pGnAisJOkrUgJ73FSK0IiYpKkK0mNJOYCR0TEvDyfI4EbSY9tujAiJpVdOTMzW/o1TVQRcciSzDgiPtug+IIuxj8NOK1B+fWkJvFmZtYLdXuPKle7fRUYVhzfr/kwM7N2KNOY4hrSldC1wFutDcfMzGxBZRLV6xFxdssjMTMza6BMojpL0onAn4A3aoURcU/LojIzM8vKJKr3Ap8HdmF+1V/kfjMzs5Yqk6g+AWxYfNWHmZlZu5R5hNJ9QL9WB2JmZtZImSuqdYGHJY1nwXtUbp5uZmYtVyZRndjyKMzMzJoo8z6q29oRiJmZWSNlnkzxMvNfrbEC0Bd4JSLWaGVgZmZmUO6KavViv6T9mP/2XTMzs5Yq0+pvARFxDf4fKjMza5MyVX+fLPQuB4ygyVt2zczMelqZVn/F91LNJb1Hat+WRGNmZlanzD2qJXovlZmZ2ZLo6lX03+5iuoiIU1sQj5mZ2QK6uqJ6pUHZqsChwDsAJyozM2u5pq3+IuKHtQ8wGlgZOAS4HNiwuxlLulDSTEkPFsrWlnSTpEfz37VyuSSdLWmKpPslbVOYZmQe/1FJI5dgXc3MbCnUZfP0nFi+A9xPuvraJiKOjYiZJeZ9EbBHXdlxwM0RMRy4OfcD7AkMz59RwLm15ZMe4bQ96X+3TqwlNzMz6x2aJipJPwDGAy8D742IkyLihbIzjojbgefrivcFLs7dFwP7FcovieROoJ+kgcDuwE0R8Xxe9k0snPzMzGwZ1tUV1THAIOAEYLqkl/LnZUkvLeby1o2IGQD57zq5fDDwZGG8abmsWbmZmfUSTRtTRMQiP7ViCahRCF2ULzwDaRSp2pD11luv5yIzM7OOamcyAngmV+mR/9budU0DhhbGGwJM76J8IRExOiJGRMSIAQMG9HjgZmbWGe1OVGOAWsu9kcDvC+UH5dZ/OwCzc9XgjcBuktbKjSh2y2VmZtZLlHmE0mKRdBmwE9Bf0jRS673TgSslHQr8E/h0Hv16YC9gCvAqqRk8EfG8pFNJjToATomI+gYaZma2DGtZooqIzzYZtGuDcQM4osl8LgQu7MHQzMxsKdLuqj8zM7NF4kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV5kRlZmaV1pFEJelxSQ9ImihpQi5bW9JNkh7Nf9fK5ZJ0tqQpku6XtE0nYjYzs87o5BXVzhGxVUSMyP3HATdHxHDg5twPsCcwPH9GAee2PVIzM+uYKlX97QtcnLsvBvYrlF8SyZ1AP0kDOxGgmZm1X6cSVQCClJhCAAALaUlEQVR/knS3pFG5bN2ImAGQ/66TywcDTxamnZbLFiBplKQJkibMmjWrhaGbmVk7Ld+h5X4gIqZLWge4SdLDXYyrBmWxUEHEaGA0wIgRIxYabmZmS6eOXFFFxPT8dybwO2A74JlalV7+OzOPPg0YWph8CDC9fdGamVkntT1RSVpV0uq1bmA34EFgDDAyjzYS+H3uHgMclFv/7QDMrlURmpnZsq8TVX/rAr+TVFv+ryPij5LGA1dKOhT4J/DpPP71wF7AFOBV4JD2h2xmZp3S9kQVEVOBLRuUPwfs2qA8gCPaEJqZmVVQlZqnm5mZLcSJyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKs2JyszMKq1Tr6K3Chp23B9avozHT9+75csws2WLE1XF9NZk0VvX28y650RlvV4nk2RvXbbZonCiMrNexQl66eNEZWZt11uTRW9d7yW11LT6k7SHpEckTZF0XKfjMTOz9lgqEpWkPsBPgT2BzYDPStqss1GZmVk7LBWJCtgOmBIRUyPiTeByYN8Ox2RmZm2giOh0DN2StD+wR0R8Mfd/Htg+Io4sjDMKGJV7NwEeWcTF9Aee7YFwe1pV4wLHtrgc26KralywbMW2fkQMaFUwi2tpaUyhBmULZNiIGA2MXuwFSBMiYsTiTt8qVY0LHNvicmyLrqpxgWNrh6Wl6m8aMLTQPwSY3qFYzMysjZaWRDUeGC5pA0krAAcAYzock5mZtcFSUfUXEXMlHQncCPQBLoyIST28mMWuNmyxqsYFjm1xObZFV9W4wLG13FLRmMLMzHqvpaXqz8zMeiknKjMzq7Relai6ewyTpBUlXZGH3yVpWIViO1jSLEkT8+eLbYrrQkkzJT3YZLgknZ3jvl/SNu2Iq2RsO0maXdhm325jbEMl3SppsqRJkr7WYJy2b7uScXVku0laSdI4Sffl2E5uME5HjtGSsXXkGC0sv4+keyVd12BYx85tPSIiesWH1AjjH8CGwArAfcBmdeMcDpyXuw8ArqhQbAcDP+nAdtsR2AZ4sMnwvYAbSP/rtgNwV4Vi2wm4rkP720Bgm9y9OvD3Bt9p27ddybg6st3ydlgtd/cF7gJ2qBunU8domdg6cowWlv+fwK8bfXed2m499elNV1RlHsO0L3Bx7r4K2FVSo3827kRsHRERtwPPdzHKvsAlkdwJ9JM0sCKxdUxEzIiIe3L3y8BkYHDdaG3fdiXj6oi8Hebk3r75U9/aqyPHaMnYOkbSEGBv4OdNRunUua1H9KZENRh4stA/jYUP0LfHiYi5wGzgHRWJDeBTuYroKklDGwzvhLKxd8r7c3XNDZI270QAuZpla9Kv8KKObrsu4oIObbdcfTURmAncFBFNt1mbj9EysUHnjtEzgW8AbzUZ3rHt1hN6U6Lq9jFMJcdphTLLvRYYFhFbAH9m/q+jTuvUNivjHtKzy7YEfgxc0+4AJK0G/BY4OiJeqh/cYJK2bLtu4urYdouIeRGxFenpM9tJek/dKB3bZiVi68gxKuljwMyIuLur0RqUVeU47VZvSlRlHsP09jiSlgfWpD1VS93GFhHPRcQbufdnwPvaEFcZlX28VUS8VKuuiYjrgb6S+rdr+ZL6kpLBryLi6gajdGTbdRdXp7dbXu6LwFhgj7pBnTpGu42tg8foB4B9JD1Oum2wi6RL68bp+HZbEr0pUZV5DNMYYGTu3h+4JfLdx07HVnfvYh/SvYUqGAMclFuw7QDMjogZnQ4KQNI7a/XwkrYj7e/PtWnZAi4AJkfEj5qM1vZtVyauTm03SQMk9cvdKwMfAR6uG60jx2iZ2Dp1jEbE8RExJCKGkc4dt0TEgXWjderc1iOWikco9YRo8hgmSacAEyJiDOkA/qWkKaRfGwdUKLajJO0DzM2xHdyO2CRdRmoF1l/SNOBE0o1kIuI84HpS67UpwKvAIe2Iq2Rs+wNfkTQXeA04oI0H5weAzwMP5PsaAN8E1ivE14ltVyauTm23gcDFSi9KXQ64MiKuq8IxWjK2jhyjzVRku/UIP0LJzMwqrTdV/ZmZ2VLIicrMzCrNicrMzCrNicrMzCrNicrMzCrNicraStK8/GTpByX9RtIqHYrj6E4tu6dI6ifp8EL/IElXdTKmZiTN6X4ss8acqKzdXouIrSLiPcCbwGFlJ8z/w9JTjgaW6kQF9CM9FRuAiJgeEft3MJ6F5H9m9nnGloh3IOukvwAbA0g6UOl9PxMlnV9LSpLmSDpF0l2kB6VuK+mv+YGp4yStnh8W+gNJ4/MDQb+cp91J0tj8gNCHJf0qnziPAgYBt0q6NY97rqQJqnvXkKS98rR3KL076rpcvqrS+7DGK70DaKGn3edl/UTSQ5L+IOl6SfvnYY8rP5ZI0ghJY7uar6TNC9vnfknDgdOBjXLZDyQNU343l9L7k34h6YE8n51z+cGSrpb0R0mPSvp+oy8mx/ddSX/L22UbSTdK+oekw/I4q0m6WdI9eTm1WIcpve/qHNJzA4cW5ts/z3PvxdhfrLdq1/tE/PEnIgDm5L/LA78HvgK8m/RAz7552DnAQbk7gH/L3SsAU4Ftc/8aeT6jgBNy2YrABGAD0lMrZpOeobcc8Dfgg3m8x4H+hbjWzn/7kJ7jtgWwEumJ0xvkYZeR3/UDfBc4MHf3I73XadW6df0kcFOe5yDgRWD/+uUDI4CxXc2X9HDYzxW2w8rAMArv4ir2A8cAv8jdmwL/zOtzcN6Ga+b+J4ChDb6nx4Gv5O4zgPtJ768aQHoAau07XCN39yc9YUM5jrcovK8JmAOsS3pS+0c7vR/6s3R9es0jlKwyVtb8R/f8hfRol1GkB3iOV3rE3MqkVykAzCM9QBVgE2BGRIyH9PBUAEm7AVvUrlZIJ+HhpKrFcRExLY83kXQSvaNBXP8maRTp5DsQ2IyU3KZGxGN5nMtyrAC7kR4E+l+5fyXSY4iKz3fbEbgsIuYB0yXdUmL7NJvv34BvKb136OqIeFRdv07og6TkRkQ8LOkJ4F152M0RMRtA0kPA+iz4upGa2vMmHyC9NPBl4GVJrys99+4V4LuSdiQlpsGkZATwRKR3bNX0BW4GjoiI20psB7O3OVFZu70W6VUJb1M6414cEcc3GP/1fKKH9Gu90TO/BHw1Im6sm+9OwBuFonk02OclbQD8F+lK7QVJF5ESRFeZQMCnIuKRLsahSbyQngdXq3pfqcR8J+fqz72BG5Vecz61m/ia6Xab1I33Vt00b+VpPke6wnpfRPxL6endtXV5pW5ec4G7gd0BJypbJL5HZVVwM7C/pHUAJK0taf0G4z0MDJK0bR5vdaVXFtxIeohq31z+LkmrdrPMl0lVWZCqEF8BZktaF9izsLwNlV4wCPCZwvQ3Al/NSRZJWzdYxu3AAfke2kBg58Kwx5n/GohPdTdfSRuSru7OJl3pbFG3Do2W/bk87btIV2XdJdVFtSapGvBf+R5Yo++sJoAvAJtKOq6H47BlnBOVdVxEPAScAPxJ0v2k+zoLvZI9It4kJYsfS7ovj7cS6fXbDwH35MYE59N9bcFo4AZJt0bEfcC9wCTgQuD/8vJeI7Wq+6OkO4BnSPe8AE4lVWfdn5d5aoNl/A54lFR1di4LXkmcDJwl6S+kq5qaZvP9DPBgrr7clPQK++eA/1Nq6v+DumWfA/SR9ABwBXBwzH9XUk/5FTBC0gRSUqx/JccC8pXxAcDOKjSrN+uOn55u1gVJq0XEnHyF81Pg0Yg4YzHndRGpMUYl/9fJrKp8RWXWtS/lq5hJpKqu8zscj1mv4ysqMzOrNF9RmZlZpTlRmZlZpTlRmZlZpTlRmZlZpTlRmZlZpf1/eGktqa7DXCcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3f68908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 0]['percentage question mark'], bins=10, range=(0, 4), rwidth=0.8)\n",
    "plt.xlabel('Percentage question mark')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage question mark of non-spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Percentage exclamation mark"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>is spam</th>\n",
       "      <th>text</th>\n",
       "      <th>percentage digits</th>\n",
       "      <th>percentage question mark</th>\n",
       "      <th>percentage exclamation mark</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>16.129032</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   is spam                                               text  \\\n",
       "1        0  Go until jurong point, crazy.. Available only ...   \n",
       "2        0                      Ok lar... Joking wif u oni...   \n",
       "3        1  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "4        0  U dun say so early hor... U c already then say...   \n",
       "5        0  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "   percentage digits  percentage question mark  percentage exclamation mark  \n",
       "1           0.000000                       0.0                          0.0  \n",
       "2           0.000000                       0.0                          0.0  \n",
       "3          16.129032                       0.0                          0.0  \n",
       "4           0.000000                       0.0                          0.0  \n",
       "5           0.000000                       0.0                          0.0  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def percentage_exclamation_mark(nnn):\n",
    "    counter = 0\n",
    "    for i in nnn:\n",
    "        if i == '!':\n",
    "            counter += 1\n",
    "    return (counter / len(nnn)) * 100\n",
    "\n",
    "df['percentage exclamation mark'] = df['text'].apply(percentage_exclamation_mark)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3f5a518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 1]['percentage exclamation mark'], bins=10, range=(0, 3), rwidth=0.8)\n",
    "plt.xlabel('Percentage exclamation mark')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage exclamation mark of spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3e7e438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 0]['percentage exclamation mark'], bins=10, range=(0, 3), rwidth=0.8)\n",
    "plt.xlabel('Percentage exclamation mark')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage exclamation mark of non-spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Percentage caps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>is spam</th>\n",
       "      <th>text</th>\n",
       "      <th>percentage digits</th>\n",
       "      <th>percentage question mark</th>\n",
       "      <th>percentage exclamation mark</th>\n",
       "      <th>percentage caps</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.702703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.896552</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>16.129032</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.451613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.081633</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.278689</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   is spam                                               text  \\\n",
       "1        0  Go until jurong point, crazy.. Available only ...   \n",
       "2        0                      Ok lar... Joking wif u oni...   \n",
       "3        1  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "4        0  U dun say so early hor... U c already then say...   \n",
       "5        0  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "   percentage digits  percentage question mark  percentage exclamation mark  \\\n",
       "1           0.000000                       0.0                          0.0   \n",
       "2           0.000000                       0.0                          0.0   \n",
       "3          16.129032                       0.0                          0.0   \n",
       "4           0.000000                       0.0                          0.0   \n",
       "5           0.000000                       0.0                          0.0   \n",
       "\n",
       "   percentage caps  \n",
       "1         2.702703  \n",
       "2         6.896552  \n",
       "3         6.451613  \n",
       "4         4.081633  \n",
       "5         3.278689  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def percentage_caps(nnn):\n",
    "    counter = 0\n",
    "    for i in nnn:\n",
    "        if i.isupper():\n",
    "            counter += 1\n",
    "    return (counter / len(nnn)) * 100\n",
    "\n",
    "df['percentage caps'] = df['text'].apply(percentage_caps)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3f3ef60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 1]['percentage caps'], bins=10, range=(0, 40), rwidth=0.8)\n",
    "plt.xlabel('Percentage caps')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage caps of spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3d7c128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 0]['percentage caps'], bins=10, range=(0, 40), rwidth=0.8)\n",
    "plt.xlabel('Percentage caps')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of percentage caps of non-spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>is spam</th>\n",
       "      <th>text</th>\n",
       "      <th>percentage digits</th>\n",
       "      <th>percentage question mark</th>\n",
       "      <th>percentage exclamation mark</th>\n",
       "      <th>percentage caps</th>\n",
       "      <th>length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.702703</td>\n",
       "      <td>111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.896552</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>16.129032</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.451613</td>\n",
       "      <td>155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.081633</td>\n",
       "      <td>49</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.278689</td>\n",
       "      <td>61</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   is spam                                               text  \\\n",
       "1        0  Go until jurong point, crazy.. Available only ...   \n",
       "2        0                      Ok lar... Joking wif u oni...   \n",
       "3        1  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "4        0  U dun say so early hor... U c already then say...   \n",
       "5        0  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "   percentage digits  percentage question mark  percentage exclamation mark  \\\n",
       "1           0.000000                       0.0                          0.0   \n",
       "2           0.000000                       0.0                          0.0   \n",
       "3          16.129032                       0.0                          0.0   \n",
       "4           0.000000                       0.0                          0.0   \n",
       "5           0.000000                       0.0                          0.0   \n",
       "\n",
       "   percentage caps  length  \n",
       "1         2.702703     111  \n",
       "2         6.896552      29  \n",
       "3         6.451613     155  \n",
       "4         4.081633      49  \n",
       "5         3.278689      61  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['length'] = df['text'].apply(len)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3d58160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 1]['length'], bins=10, range=(0, 200), rwidth=0.8)\n",
    "plt.xlabel('Length of message')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of length of spam messages.')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3d13630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 0]['length'], bins=10, range=(0, 200), rwidth=0.8)\n",
    "plt.xlabel('Length of message')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of length of non-spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TASK 3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Your task is to analyse the data, either manually or by Pandas' functions such as `groupby` and `describe` and add atleast one feature. The features created should be supported by some statistic, or some plot. \n",
    "\n",
    "Treat this task very seriously. Students whose features are most creative (and effective) will be given a special mention."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# YOUR CODE STARTS HERE\n",
    "### \"...\" are mostly sent during a chat or sms conversation\n",
    "def countofunderstanding(nnn):\n",
    "    counter = 0\n",
    "    simul=0\n",
    "    for i in nnn:\n",
    "        if i=='.':\n",
    "            simul+=1\n",
    "        else:\n",
    "            simul=0\n",
    "        if simul>=2:\n",
    "            counter += 1\n",
    "            simul=0\n",
    "    return counter\n",
    "df['countofunderstanding']= df['text'].apply(countofunderstanding)\n",
    "\n",
    "# YOUR CODE ENDS HERE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3c93eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 1]['countofunderstanding'], bins=30, range=(0, 10), rwidth=0.8)\n",
    "plt.xlabel('Number of ...')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of No. of \"...\" in spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3c5a358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 0]['countofunderstanding'], bins=30, range=(0, 10), rwidth=0.8)\n",
    "plt.xlabel('Number of ...')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of No. of \"...\" in non-spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def isalpha(word):\n",
    "    word = word.replace('.', '')\n",
    "    return word.isalpha()\n",
    "\n",
    "def clean_sms(text):\n",
    "    text = text.lower()\n",
    "    return (' '.join(filter(lambda x: isalpha(x) and x not in stopwords.words(\"english\"), text.split()))).replace('.', '').split()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "file=open('spam_keywords.txt','r')\n",
    "str=file.read()\n",
    "str=str.lower().split('\\n')\n",
    "def isspamkey(nnn):\n",
    "    count=0\n",
    "    for i in clean_sms(nnn):\n",
    "        if i in str:\n",
    "            count+=1\n",
    "    return count\n",
    "df['keyword']=df['text'].apply(isspamkey)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3c170f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 1]['keyword'], bins=30, range=(0, 10), rwidth=0.8)\n",
    "plt.xlabel('Number of ...')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of No. of \"...\" in spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd3d3d25160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(df[df['is spam'] == 0]['keyword'], bins=30, range=(0, 10), rwidth=0.8)\n",
    "plt.xlabel('Number of ...')\n",
    "plt.ylabel('Number of messages')\n",
    "plt.title('Distribution of No. of \"...\" in non-spam messages.')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating Features from the text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Uptil now, we have created additional features from the text. But we have not represented the text itself.   \n",
    "\n",
    "Remember that machine learning algorithms such as Logistic Regression work by multiplying the feature values by some (learnable) weights. Clealy, we can't use text as it as feature. We need to convert into numbers, a form machine learning algorithm can deal with.     "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Bag of Words"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "    How do we convert them to some kind of numeric representation for machine learning? One common approach is called a Bag of Words. The Bag of Words model learns a vocabulary from all of the documents, then models each document by counting the number of times each word appears. For example, consider the following two sentences:\n",
    "\n",
    "    Sentence 1: \"The cat sat on the hat\"\n",
    "\n",
    "    Sentence 2: \"The dog ate the cat and the hat\"\n",
    "\n",
    "    From these two sentences, our vocabulary is as follows:\n",
    "\n",
    "    { the, cat, sat, on, hat, dog, ate, and }\n",
    "\n",
    "    To get our bags of words, we count the number of times each word occurs in each sentence. In Sentence 1, \"the\" appears twice, and \"cat\", \"sat\", \"on\", and \"hat\" each appear once, so the feature vector for Sentence 1 is:\n",
    "\n",
    "    { the, cat, sat, on, hat, dog, ate, and }\n",
    "\n",
    "    Sentence 1: { 2, 1, 1, 1, 1, 0, 0, 0 }\n",
    "\n",
    "    Similarly, the features for Sentence 2 are: { 3, 1, 0, 0, 1, 1, 1, 1}\n",
    "    \n",
    "More on Bag of Words, [here](https://www.kaggle.com/c/word2vec-nlp-tutorial#part-1-for-beginners-bag-of-words), [here](https://machinelearningmastery.com/gentle-introduction-bag-words-model/) and [here](http://scikit-learn.org/stable/modules/feature_extraction.html#text-feature-extraction).\n",
    "\n",
    "It is imperative that you understand Bag of Words before moving on.\n",
    "\n",
    "Bag of Words is not fact the only (and the best) way to represent text. It suffers from problems such as creating too many features, not being able to represent similarities between similar words, and not being able to take into account the sequence in which words appear.  \n",
    "After learning Deep Learning, you will come to know of other methods to represent text such as using Word vectors and LSTMs."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cleaning the text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we clean the text by removing non-alphabetic words and [stopwords](http://text-analytics101.rxnlp.com/2014/10/all-about-stop-words-for-text-mining.html) from the text. Stop Words are words that most commonly appear in the language, for example, 'the', 'and', etc. They don't add useful information to the text so they can be removed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def isalpha(word):\n",
    "    word = word.replace('.', '')\n",
    "    return word.isalpha()\n",
    "\n",
    "def clean_sms(text):\n",
    "    text = text.lower()\n",
    "    return (' '.join(filter(lambda x: isalpha(x) and x not in stopwords.words(\"english\"), text.split()))).replace('.', '').split()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will be using `CountVectorizer` to convert the SMSes to a matrix of token counts. You should try and fiddle with the arguments to `CountVectorizer`"
   ]
  },
  {
   "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>is spam</th>\n",
       "      <th>text</th>\n",
       "      <th>percentage digits</th>\n",
       "      <th>percentage question mark</th>\n",
       "      <th>percentage exclamation mark</th>\n",
       "      <th>percentage caps</th>\n",
       "      <th>length</th>\n",
       "      <th>countofunderstanding</th>\n",
       "      <th>keyword</th>\n",
       "      <th>abiola</th>\n",
       "      <th>...</th>\n",
       "      <th>yo</th>\n",
       "      <th>yoga</th>\n",
       "      <th>you</th>\n",
       "      <th>yours</th>\n",
       "      <th>yr</th>\n",
       "      <th>yrs</th>\n",
       "      <th>yun</th>\n",
       "      <th>yup</th>\n",
       "      <th>zed</th>\n",
       "      <th>ìï</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.702703</td>\n",
       "      <td>111</td>\n",
       "      <td>3</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>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.896552</td>\n",
       "      <td>29</td>\n",
       "      <td>2</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>2 rows × 1461 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   is spam                                               text  \\\n",
       "1        0  Go until jurong point, crazy.. Available only ...   \n",
       "2        0                      Ok lar... Joking wif u oni...   \n",
       "\n",
       "   percentage digits  percentage question mark  percentage exclamation mark  \\\n",
       "1                0.0                       0.0                          0.0   \n",
       "2                0.0                       0.0                          0.0   \n",
       "\n",
       "   percentage caps  length  countofunderstanding  keyword  abiola ...  yo  \\\n",
       "1         2.702703     111                     3        0       0 ...   0   \n",
       "2         6.896552      29                     2        0       0 ...   0   \n",
       "\n",
       "   yoga  you  yours  yr  yrs  yun  yup  zed  ìï  \n",
       "1     0    0      0   0    0    0    0    0   0  \n",
       "2     0    0      0   0    0    0    0    0   0  \n",
       "\n",
       "[2 rows x 1461 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv = CountVectorizer(strip_accents='ascii', min_df=5, analyzer=clean_sms)\n",
    "df = pd.concat([df, pd.DataFrame(cv.fit_transform(df['text']).todense(), columns=cv.get_feature_names(), index=np.arange(1, cv.fit_transform(df['text']).todense().shape[0] + 1))], axis=1)\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is how our final feature vector (plus the target vector) looks like."
   ]
  },
  {
   "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>percentage digits</th>\n",
       "      <th>percentage question mark</th>\n",
       "      <th>percentage exclamation mark</th>\n",
       "      <th>percentage caps</th>\n",
       "      <th>length</th>\n",
       "      <th>countofunderstanding</th>\n",
       "      <th>keyword</th>\n",
       "      <th>abiola</th>\n",
       "      <th>able</th>\n",
       "      <th>about</th>\n",
       "      <th>...</th>\n",
       "      <th>yo</th>\n",
       "      <th>yoga</th>\n",
       "      <th>you</th>\n",
       "      <th>yours</th>\n",
       "      <th>yr</th>\n",
       "      <th>yrs</th>\n",
       "      <th>yun</th>\n",
       "      <th>yup</th>\n",
       "      <th>zed</th>\n",
       "      <th>ìï</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>is spam</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.420082</td>\n",
       "      <td>0.544954</td>\n",
       "      <td>0.233188</td>\n",
       "      <td>5.952566</td>\n",
       "      <td>71.023627</td>\n",
       "      <td>0.408912</td>\n",
       "      <td>0.060725</td>\n",
       "      <td>0.00228</td>\n",
       "      <td>0.005389</td>\n",
       "      <td>0.001036</td>\n",
       "      <td>...</td>\n",
       "      <td>0.005389</td>\n",
       "      <td>0.001244</td>\n",
       "      <td>0.017409</td>\n",
       "      <td>0.000622</td>\n",
       "      <td>0.000622</td>\n",
       "      <td>0.001036</td>\n",
       "      <td>0.001036</td>\n",
       "      <td>0.008497</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.010777</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>11.579405</td>\n",
       "      <td>0.169521</td>\n",
       "      <td>0.524716</td>\n",
       "      <td>11.061228</td>\n",
       "      <td>138.866131</td>\n",
       "      <td>0.033467</td>\n",
       "      <td>0.935743</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001339</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.010710</td>\n",
       "      <td>0.010710</td>\n",
       "      <td>0.014726</td>\n",
       "      <td>0.004016</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.006693</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 1459 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         percentage digits  percentage question mark  \\\n",
       "is spam                                                \n",
       "0                 0.420082                  0.544954   \n",
       "1                11.579405                  0.169521   \n",
       "\n",
       "         percentage exclamation mark  percentage caps      length  \\\n",
       "is spam                                                             \n",
       "0                           0.233188         5.952566   71.023627   \n",
       "1                           0.524716        11.061228  138.866131   \n",
       "\n",
       "         countofunderstanding   keyword   abiola      able     about  \\\n",
       "is spam                                                                \n",
       "0                    0.408912  0.060725  0.00228  0.005389  0.001036   \n",
       "1                    0.033467  0.935743  0.00000  0.000000  0.000000   \n",
       "\n",
       "           ...           yo      yoga       you     yours        yr       yrs  \\\n",
       "is spam    ...                                                                  \n",
       "0          ...     0.005389  0.001244  0.017409  0.000622  0.000622  0.001036   \n",
       "1          ...     0.001339  0.000000  0.010710  0.010710  0.014726  0.004016   \n",
       "\n",
       "              yun       yup       zed        ìï  \n",
       "is spam                                          \n",
       "0        0.001036  0.008497  0.000000  0.010777  \n",
       "1        0.000000  0.000000  0.006693  0.000000  \n",
       "\n",
       "[2 rows x 1459 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('is spam').mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The table above shows mean values for all the features grouped by whether the text is spam or ham.  \n",
    "Clearly, the values are very different for the features we manually created. That means we have created good features."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training model and calculating accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TASK 4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Extract `X`, the feature vectors by removing the label (`is spam` column) and the text (`text` column) from the `df` DataFrame.  \n",
    "Similarly extract the target vectors, by assigning `Y` to the `is spam` column."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# YOUR CODE STARTS HERE\n",
    "X=df.drop(columns=['text','is spam'])\n",
    "y=df['is spam']\n",
    "# YOUR CODE ENDS HERE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X.shape: (5572, 1458)\n",
      "y.shape: (5572,)\n"
     ]
    }
   ],
   "source": [
    "print(\"X.shape:\", X.shape)\n",
    "print(\"y.shape:\", y.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we divide the dataset into 3 parts, `train`, `cv` and `test`. We train our model on the training set, tune the hyperparameters (regularisation factor, which algorithm to use) on the validation set and finally report the our results on the test set.\n",
    "\n",
    "In this case we keep 60% of the data in training set, 20% of the data in validation set, and 20% of the data in test set. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Splitting the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train dataset size:  3343\n",
      "CV size:  1114\n",
      "Test size:  1115\n"
     ]
    }
   ],
   "source": [
    "X_train, X_other, y_train, y_other = train_test_split(X, y, test_size=0.4, random_state=42)\n",
    "X_cv, X_test, y_cv, y_test = train_test_split(X_other, y_other, test_size=0.5, random_state=42)\n",
    "\n",
    "print(\"Train dataset size: \", X_train.shape[0])\n",
    "print(\"CV size: \", X_cv.shape[0])\n",
    "print(\"Test size: \", X_test.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we try out different models and fiddle around with the hyperparameters, seeing their accuracy on the validation set."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Logistic Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now will use the Logistic Regression class from the scikit-learn library to create a classifier object and train it on our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  1.25 seconds\n"
     ]
    }
   ],
   "source": [
    "clf1 = LogisticRegression(C=2.6,solver='lbfgs')\n",
    "from time import time as t\n",
    "t0=t()\n",
    "clf1 = clf1.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fiddle around the paramaters while creating the classifier object, particularly with `C` - the regularisation strength, `class_weight` - a useful paramter when dealing with skewed class distribution like this. You can change the regularisation strength by observing how it is performing on the training set and the validation set."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will be using f1-score to evaluate the model. F1-score is used when the class distribution is skewed (like in this example).   \n",
    "More about F1-Score [here](https://machinelearningmastery.com/classification-accuracy-is-not-enough-more-performance-measures-you-can-use/). Do read this before moving on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.99      1.00      0.99       955\n",
      "       spam       0.99      0.95      0.97       159\n",
      "\n",
      "avg / total       0.99      0.99      0.99      1114\n",
      "\n",
      "F1 score: 0.9679487179487181\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf1.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf1.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Naive Bayes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similarly fiddle around with the hyper parameters of this model. You can find the documentation on the scikit-learn website."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  0.034 seconds\n"
     ]
    }
   ],
   "source": [
    "clf2 = MultinomialNB(alpha=0.00023)\n",
    "t0=t()\n",
    "clf2 = clf2.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.99      0.99      0.99       955\n",
      "       spam       0.92      0.92      0.92       159\n",
      "\n",
      "avg / total       0.98      0.98      0.98      1114\n",
      "\n",
      "F1 score: 0.9182389937106918\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf2.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf2.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  1.063 seconds\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import AdaBoostClassifier as ABC\n",
    "clf5 = ABC()\n",
    "t0=t()\n",
    "clf5 = clf5.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.98      0.99      0.99       955\n",
      "       spam       0.95      0.90      0.92       159\n",
      "\n",
      "avg / total       0.98      0.98      0.98      1114\n",
      "\n",
      "F1 score: 0.9225806451612903\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf5.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf5.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  0.186 seconds\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier as RFC\n",
    "clf6 = RFC(min_samples_split=19)\n",
    "t0=t()\n",
    "clf6 = clf6.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.98      1.00      0.99       955\n",
      "       spam       0.99      0.86      0.92       159\n",
      "\n",
      "avg / total       0.98      0.98      0.98      1114\n",
      "\n",
      "F1 score: 0.9158249158249159\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf6.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf6.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  0.372 seconds\n"
     ]
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier as DTC\n",
    "clf7 = DTC(min_samples_split=16)\n",
    "t0=t()\n",
    "clf7 = clf7.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.99      0.99      0.99       955\n",
      "       spam       0.95      0.91      0.93       159\n",
      "\n",
      "avg / total       0.98      0.98      0.98      1114\n",
      "\n",
      "F1 score: 0.9294871794871795\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf7.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf7.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  0.286 seconds\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier as KNC\n",
    "clf8 = KNC(n_neighbors=7)\n",
    "t0=t()\n",
    "clf8 = clf8.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.98      0.99      0.99       955\n",
      "       spam       0.95      0.91      0.93       159\n",
      "\n",
      "avg / total       0.98      0.98      0.98      1114\n",
      "\n",
      "F1 score: 0.9260450160771704\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf8.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf8.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  12.479 seconds\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neural_network import MLPClassifier as MLPC\n",
    "clf9 = MLPC(solver='lbfgs')\n",
    "t0=t()\n",
    "clf9 = clf9.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.99      1.00      0.99       955\n",
      "       spam       0.98      0.95      0.96       159\n",
      "\n",
      "avg / total       0.99      0.99      0.99      1114\n",
      "\n",
      "F1 score: 0.964856230031949\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf9.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf9.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### TASK 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use [SVM Classifier](http://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html) instead of Logistic Regression and Naive Bayes. \n",
    "Print out the classification report and F1 Score as done above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  5.018 seconds\n"
     ]
    }
   ],
   "source": [
    "# YOUR CODE STARTS HERE\n",
    "from sklearn.svm import SVC\n",
    "clf3 = SVC(kernel='linear',C=0.17)\n",
    "t0=t()\n",
    "clf3 = clf3.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")\n",
    "# YOUR CODE ENDS HERE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.99      1.00      0.99       955\n",
      "       spam       0.98      0.95      0.96       159\n",
      "\n",
      "avg / total       0.99      0.99      0.99      1114\n",
      "\n",
      "F1 score: 0.964856230031949\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_cv, clf3.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print('F1 score:', f1_score(y_cv, clf3.predict(X_cv)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Bonus Task"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Try using one of the most basic ensembling method the [VotingClassifier](http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.VotingClassifier.html).\n",
    "You can read more about ensembing [here](http://blog.kaggle.com/2016/12/27/a-kagglers-guide-to-model-stacking-in-practice/), and [here](https://mlwave.com/kaggle-ensembling-guide/)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time Taken=  8.318 seconds\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "        ham       0.99      1.00      1.00       955\n",
      "       spam       0.99      0.95      0.97       159\n",
      "\n",
      "avg / total       0.99      0.99      0.99      1114\n",
      "\n",
      "F1 score: 0.9710610932475885\n"
     ]
    }
   ],
   "source": [
    "# YOUR CODE STARTS HERE\n",
    "from sklearn.ensemble import VotingClassifier as VC\n",
    "clf4 = VC(estimators=[('logistic', clf1), ('multinomial', clf2),('svm',clf3),('AdaBoost',clf5),('RandomForest',clf6),('DecisionTree',clf7),('KNearestNeighbors',clf8)])\n",
    "t0=t()\n",
    "clf4 = clf4.fit(X_train, y_train)\n",
    "print(\"Time Taken= \",round(t()-t0,3),\"seconds\")\n",
    "\n",
    "print(classification_report(y_cv, clf4.predict(X_cv), target_names=['ham', 'spam']))\n",
    "print(\"F1 score:\", f1_score(y_cv, clf4.predict(X_cv)))\n",
    "# YOUR CODE ENDS HERE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training the final model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_actual = y_test\n",
    "y_predicted = clf4.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [],
   "source": [
    "true_positives = X_test[(y_actual == 1) & (y_predicted == 1)]\n",
    "true_negatives = X_test[(y_actual == 0) & (y_predicted == 0)]\n",
    "false_positives = X_test[(y_actual == 0) & (y_predicted == 1)]\n",
    "false_negatives = X_test[(y_actual == 1) & (y_predicted == 0)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision: 0.9850746268656716\n",
      "Recall: 0.9428571428571428\n",
      "F1 score: 0.9635036496350364\n"
     ]
    }
   ],
   "source": [
    "precision = true_positives.shape[0] / (true_positives.shape[0] + false_positives.shape[0])\n",
    "print(\"Precision:\", precision)\n",
    "recall = true_positives.shape[0] / (true_positives.shape[0] + false_negatives.shape[0])\n",
    "print(\"Recall:\", recall)\n",
    "f1_score = 2 * precision * recall / (precision + recall)\n",
    "print(\"F1 score:\", f1_score)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The one with the highest F1 Score will get a special mention. To improve on the F1 Score, you can engineer new features, try out more algorithms and loggically fiddle with their parameters."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Congratulations! You have successfully made a spam classifier which is correct most of the time. Isn't it amazing how your computer has learnt to distinguish between spam and non-spam messages? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Observing examples where the algorithm failed, or was successful."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "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>text</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>991</th>\n",
       "      <td>26th OF JULY</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4728</th>\n",
       "      <td>I (Career Tel) have added u as a contact on IN...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   text  text\n",
       "991                                        26th OF JULY     0\n",
       "4728  I (Career Tel) have added u as a contact on IN...     0"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[list(false_positives.index)]['text']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "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>text</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1500</th>\n",
       "      <td>SMS. ac JSco: Energy is high, but u may not kn...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3990</th>\n",
       "      <td>(Bank of Granite issues Strong-Buy) EXPLOSIVE ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>Did you hear about the new \\Divorce Barbie\\\"? ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2313</th>\n",
       "      <td>tddnewsletter@emc1.co.uk (More games from TheD...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>869</th>\n",
       "      <td>Hello. We need some posh birds and chaps to us...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>752</th>\n",
       "      <td>You have an important customer service announc...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2663</th>\n",
       "      <td>Hello darling how are you today? I would love ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2003</th>\n",
       "      <td>TheMob&gt;Yo yo yo-Here comes a new selection of ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   text  text\n",
       "1500  SMS. ac JSco: Energy is high, but u may not kn...     0\n",
       "3990  (Bank of Granite issues Strong-Buy) EXPLOSIVE ...     0\n",
       "69    Did you hear about the new \\Divorce Barbie\\\"? ...     0\n",
       "2313  tddnewsletter@emc1.co.uk (More games from TheD...     0\n",
       "869   Hello. We need some posh birds and chaps to us...     0\n",
       "752   You have an important customer service announc...     0\n",
       "2663  Hello darling how are you today? I would love ...     0\n",
       "2003  TheMob>Yo yo yo-Here comes a new selection of ...     0"
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[list(false_negatives.index)]['text']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}