Introduction_to_Practical_Machine_Learning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "MWrlNEQJWohC"
   },
   "source": [
    "\n",
    "# Introduction to Classification\n",
    "\n",
    "* Classification techniques are an essential part of machine learning and data mining applications.\n",
    "\n",
    "* Approximately 70% of problems in Data Science are classification problems. \n",
    "\n",
    "\n",
    "* In solving most of these problems, a go to algorithm is the logistic regression model.\n",
    "\n",
    "* It is useful in solving binary class problems.\n",
    "\n",
    "In this practical session, we will train a binary class problem using \n",
    "**logistic regression**.\n",
    "\n",
    "* Logistic regression is useful for various classification problems including spam detection, diabetes prediction, churning of customers etc\n",
    "\n",
    "By the end of this lab, you will be able to:\n",
    "\n",
    "* understand difference between linear regression and logistic regression\n",
    "\n",
    "* how logistic regression works?\n",
    "\n",
    "* train a logisitc regression model in Scikit-learn\n",
    " \n",
    "* evaluate a logistic regression model\n",
    "\n",
    "\n",
    "* outline the advantages and disadvantages of Logistic Regression\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "mE_8nu3UWohC"
   },
   "source": [
    "## Note:\n",
    "\n",
    "\n",
    "* Logistic regression is a type of a classification algorithm in which the outcome is binary/discrete.\n",
    "\n",
    "* Logistic regression is a special case of linear regression where the \n",
    "target variable is categorical in nature. \n",
    "\n",
    "* As such, Logistic Regression predicts the probability of occurrence of a binary event utilizing a logit function.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "4L72jMtLWohC"
   },
   "source": [
    "Given a dataset $D=\\{X_{i},y_{i}\\}^{n}_{i=1}$  as inputs with output class $y \\in \\{0,1\\}$. We would define our hypothesis $h_0$ as:\n",
    "$$\\dfrac{1}{1+ e^{-\\theta ^{T}X}} \\hspace{0.2in} \\text{where} \\hspace{0.2in} \\dfrac{1}{1+ e^{-z}} \\hspace{0.2in} \\text{is sigmoid function}$$ \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "mV7pC00_WohD"
   },
   "source": [
    "### Objective:\n",
    "\n",
    "We would be training a logistic regression model using a [dataset](https://zindi.africa/hackathons/umojahack-ghana-expresso-churn-prediction-challenge) from  Zindi on churn predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "vds2R2yaWohD",
    "outputId": "51902010-7160-487b-802f-12ee34e504e5"
   },
   "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>user_id</th>\n",
       "      <th>REGION</th>\n",
       "      <th>TENURE</th>\n",
       "      <th>MONTANT</th>\n",
       "      <th>FREQUENCE_RECH</th>\n",
       "      <th>REVENUE</th>\n",
       "      <th>ARPU_SEGMENT</th>\n",
       "      <th>FREQUENCE</th>\n",
       "      <th>DATA_VOLUME</th>\n",
       "      <th>ON_NET</th>\n",
       "      <th>ORANGE</th>\n",
       "      <th>TIGO</th>\n",
       "      <th>ZONE1</th>\n",
       "      <th>ZONE2</th>\n",
       "      <th>MRG</th>\n",
       "      <th>REGULARITY</th>\n",
       "      <th>TOP_PACK</th>\n",
       "      <th>FREQ_TOP_PACK</th>\n",
       "      <th>CHURN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>00000bfd7d50f01092811bc0c8d7b0d6fe7c3596</td>\n",
       "      <td>FATICK</td>\n",
       "      <td>K &gt; 24 month</td>\n",
       "      <td>4250.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>4251.0</td>\n",
       "      <td>1417.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>388.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NO</td>\n",
       "      <td>54</td>\n",
       "      <td>On net 200F=Unlimited _call24H</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>00000cb4a5d760de88fecb38e2f71b7bec52e834</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I 18-21 month</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NO</td>\n",
       "      <td>4</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>00001654a9d9f96303d9969d0a4a851714a4bb57</td>\n",
       "      <td>NaN</td>\n",
       "      <td>K &gt; 24 month</td>\n",
       "      <td>3600.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1020.0</td>\n",
       "      <td>340.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>90.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NO</td>\n",
       "      <td>17</td>\n",
       "      <td>On-net 1000F=10MilF;10d</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>00001dd6fa45f7ba044bd5d84937be464ce78ac2</td>\n",
       "      <td>DAKAR</td>\n",
       "      <td>K &gt; 24 month</td>\n",
       "      <td>13500.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>13502.0</td>\n",
       "      <td>4501.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>43804.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>102.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NO</td>\n",
       "      <td>62</td>\n",
       "      <td>Data:1000F=5GB,7d</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>000028d9e13a595abe061f9b58f3d76ab907850f</td>\n",
       "      <td>DAKAR</td>\n",
       "      <td>K &gt; 24 month</td>\n",
       "      <td>1000.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>985.0</td>\n",
       "      <td>328.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>39.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NO</td>\n",
       "      <td>11</td>\n",
       "      <td>Mixt 250F=Unlimited_call24H</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                    user_id  REGION         TENURE  MONTANT  \\\n",
       "0  00000bfd7d50f01092811bc0c8d7b0d6fe7c3596  FATICK   K > 24 month   4250.0   \n",
       "1  00000cb4a5d760de88fecb38e2f71b7bec52e834     NaN  I 18-21 month      NaN   \n",
       "2  00001654a9d9f96303d9969d0a4a851714a4bb57     NaN   K > 24 month   3600.0   \n",
       "3  00001dd6fa45f7ba044bd5d84937be464ce78ac2   DAKAR   K > 24 month  13500.0   \n",
       "4  000028d9e13a595abe061f9b58f3d76ab907850f   DAKAR   K > 24 month   1000.0   \n",
       "\n",
       "   FREQUENCE_RECH  REVENUE  ARPU_SEGMENT  FREQUENCE  DATA_VOLUME  ON_NET  \\\n",
       "0            15.0   4251.0        1417.0       17.0          4.0   388.0   \n",
       "1             NaN      NaN           NaN        NaN          NaN     NaN   \n",
       "2             2.0   1020.0         340.0        2.0          NaN    90.0   \n",
       "3            15.0  13502.0        4501.0       18.0      43804.0    41.0   \n",
       "4             1.0    985.0         328.0        1.0          NaN    39.0   \n",
       "\n",
       "   ORANGE  TIGO  ZONE1  ZONE2 MRG  REGULARITY                        TOP_PACK  \\\n",
       "0    46.0   1.0    1.0    2.0  NO          54  On net 200F=Unlimited _call24H   \n",
       "1     NaN   NaN    NaN    NaN  NO           4                             NaN   \n",
       "2    46.0   7.0    NaN    NaN  NO          17         On-net 1000F=10MilF;10d   \n",
       "3   102.0   2.0    NaN    NaN  NO          62               Data:1000F=5GB,7d   \n",
       "4    24.0   NaN    NaN    NaN  NO          11     Mixt 250F=Unlimited_call24H   \n",
       "\n",
       "   FREQ_TOP_PACK  CHURN  \n",
       "0            8.0      0  \n",
       "1            NaN      1  \n",
       "2            1.0      0  \n",
       "3           11.0      0  \n",
       "4            2.0      0  "
      ]
     },
     "execution_count": 1,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "df = pd.read_csv('Train .csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "jS350Kg9WohE",
    "outputId": "f1bf421e-f14a-4318-b493-b8b0395a17cb",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Variable Definitions</th>\n",
       "      <th>Unnamed: 1</th>\n",
       "      <th>Unnamed: 2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NaN</td>\n",
       "      <td>French</td>\n",
       "      <td>English</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NaN</td>\n",
       "      <td>Le dataset churn comprend 19 variables dont 15...</td>\n",
       "      <td>The churn dataset includes 19 variables includ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>user_id</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>REGION</td>\n",
       "      <td>la localité de chaque client</td>\n",
       "      <td>the location of each client</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TENURE</td>\n",
       "      <td>la durée dans le reseau</td>\n",
       "      <td>duration in the network</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Variable Definitions                                         Unnamed: 1  \\\n",
       "0                  NaN                                             French   \n",
       "1                  NaN  Le dataset churn comprend 19 variables dont 15...   \n",
       "2              user_id                                                NaN   \n",
       "3               REGION                       la localité de chaque client   \n",
       "4               TENURE                            la durée dans le reseau   \n",
       "\n",
       "                                          Unnamed: 2  \n",
       "0                                            English  \n",
       "1  The churn dataset includes 19 variables includ...  \n",
       "2                                                NaN  \n",
       "3                        the location of each client  \n",
       "4                            duration in the network  "
      ]
     },
     "execution_count": 3,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "description = pd.read_csv(\"VariableDefinitions.csv\")\n",
    "description.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "_j-eM514WohF",
    "outputId": "68413321-7414-4a87-d60b-d56802dedb0b",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "data cols are: Index(['user_id', 'REGION', 'TENURE', 'MONTANT', 'FREQUENCE_RECH', 'REVENUE',\n",
      "       'ARPU_SEGMENT', 'FREQUENCE', 'DATA_VOLUME', 'ON_NET', 'ORANGE', 'TIGO',\n",
      "       'ZONE1', 'ZONE2', 'MRG', 'REGULARITY', 'TOP_PACK', 'FREQ_TOP_PACK',\n",
      "       'CHURN'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(\"data cols are:\",df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "4AeD30z3WohF",
    "outputId": "178ce720-02f8-42fb-f310-c9e692cb90b4",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    1750062\n",
       "1     403986\n",
       "Name: CHURN, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.CHURN.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ktIY3r6SWohG",
    "outputId": "044db7e1-e812-4d0b-f67a-84ab2db11fa6",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 2154048 entries, 0 to 2154047\n",
      "Data columns (total 19 columns):\n",
      " #   Column          Dtype  \n",
      "---  ------          -----  \n",
      " 0   user_id         object \n",
      " 1   REGION          object \n",
      " 2   TENURE          object \n",
      " 3   MONTANT         float64\n",
      " 4   FREQUENCE_RECH  float64\n",
      " 5   REVENUE         float64\n",
      " 6   ARPU_SEGMENT    float64\n",
      " 7   FREQUENCE       float64\n",
      " 8   DATA_VOLUME     float64\n",
      " 9   ON_NET          float64\n",
      " 10  ORANGE          float64\n",
      " 11  TIGO            float64\n",
      " 12  ZONE1           float64\n",
      " 13  ZONE2           float64\n",
      " 14  MRG             object \n",
      " 15  REGULARITY      int64  \n",
      " 16  TOP_PACK        object \n",
      " 17  FREQ_TOP_PACK   float64\n",
      " 18  CHURN           int64  \n",
      "dtypes: float64(12), int64(2), object(5)\n",
      "memory usage: 312.2+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Ox04S8nNWohG",
    "outputId": "6e592d84-60fe-4a1e-ff34-87145ff5f000",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'REGION')"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "df['REGION'].value_counts(normalize=True).plot(kind='bar')\n",
    "plt.ylabel('counts')\n",
    "plt.xlabel('REGION')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "X7CbPn-6WohG",
    "outputId": "da5f858b-6218-4e15-b140-af24fb41da74",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'TENURE')"
      ]
     },
     "execution_count": 87,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['TENURE'].value_counts(normalize=True).plot(kind='bar')\n",
    "plt.ylabel('counts')\n",
    "plt.xlabel('TENURE')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_Stt6abZWohH"
   },
   "source": [
    "For the purposes of this lab, I will want to select a few columns to work with for now.\n",
    "\n",
    "\n",
    "**Disclaimer:** Before dropping out columns in any future project, you should have an explanation as to why that column is not important to your model training."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "_lKiFdLKWohH"
   },
   "outputs": [],
   "source": [
    "df = df[['REGION',  'REVENUE','DATA_VOLUME', 'ON_NET', 'CHURN']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "NE9QZzA-WohH",
    "outputId": "da6f9a80-3cde-4aa2-c922-07ed3c7539c4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The number of missing values in the data is: 3422455\n"
     ]
    }
   ],
   "source": [
    "print(\"The number of missing values in the data is:\", df.isnull().sum().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "fSgq5rlUWohI",
    "outputId": "0822a8ef-3950-4d9d-830b-575089c5d4d2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Names of cols with missing values: REGION          849299\n",
      "REVENUE         726048\n",
      "DATA_VOLUME    1060433\n",
      "ON_NET          786675\n",
      "CHURN                0\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(\"Names of cols with missing values:\", df.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "V_8cXOHsWohI"
   },
   "source": [
    "\n",
    "### Preprocessing Data \n",
    "\n",
    "Next step is to preprocess the data for training.\n",
    "\n",
    "First, we realize the data has so many missing values and as such we would need to handle them. We have different ways in handling missing values and it also depends on the data type.\n",
    "Additionally, we need to change all categorical columns to numerical and then we would scale the numerical columns too. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Rcl4UU6uWohI",
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "df['REGION'].fillna(df['REGION'].value_counts().index[0], inplace=True)\n",
    "df['DATA_VOLUME'].fillna((df['DATA_VOLUME'].mean()), inplace=True)\n",
    "df['ON_NET'].fillna((df['ON_NET'].mean()), inplace=True)\n",
    "df['REVENUE'].fillna((df['REVENUE'].mean()), inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "UYXsVMAJWohI",
    "outputId": "8aa00b61-5aa1-4744-99eb-5d7d2f5b91ae"
   },
   "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>REVENUE</th>\n",
       "      <th>DATA_VOLUME</th>\n",
       "      <th>ON_NET</th>\n",
       "      <th>CHURN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>REVENUE</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.239594</td>\n",
       "      <td>0.321036</td>\n",
       "      <td>-0.053808</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DATA_VOLUME</th>\n",
       "      <td>0.239594</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.006394</td>\n",
       "      <td>-0.015398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ON_NET</th>\n",
       "      <td>0.321036</td>\n",
       "      <td>-0.006394</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.026601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CHURN</th>\n",
       "      <td>-0.053808</td>\n",
       "      <td>-0.015398</td>\n",
       "      <td>-0.026601</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              REVENUE  DATA_VOLUME    ON_NET     CHURN\n",
       "REVENUE      1.000000     0.239594  0.321036 -0.053808\n",
       "DATA_VOLUME  0.239594     1.000000 -0.006394 -0.015398\n",
       "ON_NET       0.321036    -0.006394  1.000000 -0.026601\n",
       "CHURN       -0.053808    -0.015398 -0.026601  1.000000"
      ]
     },
     "execution_count": 12,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "d6W-evmAWohJ"
   },
   "source": [
    "## Machine Learning\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "7x4_X1SEWohJ"
   },
   "source": [
    "Select the columns for X and y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "uMu6rO8TWohJ"
   },
   "outputs": [],
   "source": [
    "dropcols = [ 'CHURN']\n",
    "y = df['CHURN']\n",
    "X = df.drop(columns=dropcols, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lcNju-myWohJ"
   },
   "source": [
    "Split data into train and test. \n",
    "\n",
    "## Why is this important?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "nUxTL1v5WohJ",
    "outputId": "973adb91-3f6b-4ad9-dcb2-5bc25428c1ff"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1723238, 4) (1723238,) (430810, 4) (430810,)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn.model_selection import train_test_split,cross_val_score\n",
    "\n",
    "X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2,random_state=1)\n",
    "print(X_train.shape,y_train.shape, X_test.shape, y_test.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "LQxJo3jnWohJ"
   },
   "outputs": [],
   "source": [
    "num_cols = ['REVENUE','DATA_VOLUME', 'ON_NET']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "gGNRZnfhWohJ"
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "scaler = StandardScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "eNoPZJUgWohJ",
    "outputId": "206cd8a3-2250-4bff-be6b-aaf73a955229",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/kanubala/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n",
      "/home/kanubala/anaconda3/lib/python3.7/site-packages/pandas/core/indexing.py:1734: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  isetter(loc, value[:, i].tolist())\n",
      "/home/kanubala/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:2: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
      "  \n"
     ]
    },
    {
     "ename": "IndexError",
     "evalue": "only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-18-f69c7bd501f2>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mX_train\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnum_cols\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mscaler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnum_cols\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mX_test\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnum_cols\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mscaler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_test\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnum_cols\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnum_cols\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mIndexError\u001b[0m: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices"
     ]
    }
   ],
   "source": [
    "X_train[num_cols] = scaler.fit_transform(X_train[num_cols])\n",
    "X_test[num_cols] = scaler.transform(X_test[num_cols])[num_cols] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "iPW4XQNPWohK",
    "outputId": "48f518bd-b3b2-4b63-9dfd-eec16601775f",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/kanubala/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n",
      "/home/kanubala/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:5: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \"\"\"\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "encoder = LabelEncoder()\n",
    "X_train[\"REGION\"] = encoder.fit_transform(X_train[\"REGION\"])\n",
    "\n",
    "X_test[\"REGION\"] = encoder.transform(X_test[\"REGION\"])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "DB8J3A9_WohK"
   },
   "source": [
    "## Model Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "SxII9zUMWohK"
   },
   "outputs": [],
   "source": [
    "# import the class\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "# instantiate the model (using the default parameters)\n",
    "model = LogisticRegression()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "_NG1Z8j7WohL",
    "outputId": "fdd56d99-4471-4b06-9f0d-0a9104e867dc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression()"
      ]
     },
     "execution_count": 21,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fit the model with data\n",
    "model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "cQ1rDOmzWohL"
   },
   "outputs": [],
   "source": [
    "\n",
    "# predict xtest\n",
    "y_pred=model.predict(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TM8irOdWWohL"
   },
   "source": [
    "### Model Evaluation using Confusion Matrix\n",
    "\n",
    "We can evaluate our model using any of the following:\n",
    "\n",
    " \n",
    "\n",
    "**Confusion matrix**\n",
    "This lays out correctly and incorrectly classified cases\n",
    "For the binary (two-class) case the confusion matrix is organized as follows:\n",
    "\n",
    "| | Scored Positive | Scored Negative|  \n",
    "|------|:------:|:------:| \n",
    "|**Actual Positive** | True Positive | False Negative |\n",
    "|**Actual Negative**| False Positive | True Negative |   \n",
    "\n",
    "Here the four elements in the matrix are defined as:    \n",
    "**True Positive** or **TP** are cases with positive labels which have been correctly classified as positive.     \n",
    "**True Negative** or **TN** are cases with negative labels which have been correctly classified as negative.  \n",
    "**False Positive** or **FP** are cases with negative labels which have been incorrectly classified as positive.   \n",
    "**False Negative** or **FN** are cases with positive labels which have been incorrectly classified as negative.\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "XRXR_pgOWohL",
    "outputId": "5656971e-52d0-47e9-cb48-c29e4fbc717e",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[349780,    318],\n",
       "       [ 80633,     79]])"
      ]
     },
     "execution_count": 23,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# import the metrics class\n",
    "from sklearn import metrics\n",
    "cnf_matrix = metrics.confusion_matrix(y_test, y_pred)\n",
    "cnf_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Ma9zH0GhWohL"
   },
   "source": [
    "**Accuracy**\n",
    "\n",
    "Accuracy is the fraction of cases correctly classified  over the total number of classifications. For a two-class classifier accuracy is written as:\n",
    "\n",
    "$$Accuracy = \\frac{TP+TN}{TP+FP+TN+FN}$$\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "nYVmH5WsWohL",
    "outputId": "8f4f21e6-d326-4166-a7ec-4ee073198aef"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8120958195027971\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "48SSHI74WohM"
   },
   "source": [
    "### Note:\n",
    "\n",
    "There are other evaluation metric such as:\n",
    "\n",
    "* F1 score\n",
    "* Recall\n",
    "* Precision\n",
    "* ROC"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ke9wkssvWohM"
   },
   "source": [
    "#### Advantages of Logisitic Regression\n",
    "\n",
    "* Easy to implement\n",
    "* Easily interpretable\n",
    "* Provides probability scores\n",
    "* Does not require high computational power\n",
    "\n",
    "\n",
    "### Disadvantages\n",
    "\n",
    "* Unable to handle handle a large number of categorical features/variables hence, it is vulnerable to overfitting.\n",
    "\n",
    "* Non-linear problems cannot be solved with logistic regression\n",
    "\n",
    "* Logistic regression will not perform well with independent variables that are not correlated to the target variable and are very similar or correlated to each other.\n",
    "\n",
    "#### Conclusion\n",
    "\n",
    "In this pratical session, we were able covered:\n",
    "\n",
    "* details about logistic regression\n",
    "\n",
    "* encoding categorical features\n",
    "\n",
    "* scaling numerical features\n",
    "\n",
    "* evaluating logisitc regression model\n",
    "\n",
    "\n",
    "Hopefully, you can now utilize the Logistic Regression technique to analyze your Zindi project. Feel free to explore other complex models for that project.\n",
    "\n",
    "\n",
    "# I wish you all the best of luck !!!"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "Introduction to Practical Machine Learning.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}