Majority Voting.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "no display found. Using non-interactive Agg backend\n"
     ]
    }
   ],
   "source": [
    "# Source code used to predict the heat classification from the file 6_25 Heat Index Sample.xlxs\n",
    "# Jalaluddin Qureshi, George Brown College, AY 2019-21, Summer 2020.\n",
    "\n",
    "# These command are being run on the Jupyter Notebook environment, which requires download and installation\n",
    "# of Anaconda (https://docs.anaconda.com/anaconda/install/). Once installed, a short introduction on using \n",
    "# Jupyter notebook can be found online such as this one, https://www.youtube.com/watch?v=HW29067qVWk.\n",
    "\n",
    "# In this cell we are importing the relevant training models which are needed to train our predictive model.\n",
    "# While running this cell for the first time there is a possibiity that some of the module may require\n",
    "# installation, the installation command is of this form $ pip install -U module_name (I tend google \n",
    "# the full actual installation command). The installation command needs to be run on anaconda command promopt.\n",
    "# Anaconda command window comes with Anaconda installation, it does not requires any separate installation.\n",
    "#\n",
    "\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from dmba import classificationSummary\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.ensemble import VotingClassifier, RandomForestClassifier, GradientBoostingClassifier \n",
    "\n",
    "from sklearn import svm #Support Vector Machine\n",
    "from sklearn.neural_network import MLPClassifier # Neural Net\n",
    "from sklearn.linear_model import LogisticRegression #Logistic Regression\n",
    "from sklearn.model_selection import GridSearchCV #Needed for Decision Tree\n",
    "from sklearn.tree import DecisionTreeClassifier #Needed for Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# It is assumed that the data file should be in the same folder in which the jupyter notebook has been opened-saved. \n",
    "# If it is not in the same folder, then that would also require providing the full path.\n",
    "\n",
    "# The data columns which are of interest are listed in the predictor variable. Other columns may be added\n",
    "# or removed. The s.strip() replaces any empty space with an underscore, e.g. \"Total Value\" becomes \"Total_Value\".\n",
    "# So the name of the variables which are added in predictor must follow this convention.\n",
    "# I had also shortened some variable name manually on Excel for simplicity, \n",
    "# e.g. \"Est # Historic Customers for product family (2015-Present)\" is relabbeled as \"historic\", and a new variable\n",
    "# \"Time_EOL\" is generated as follow: Time_EOL=2020-EOL, to determine the number of years since the compent became\n",
    "# obselete.\n",
    "\n",
    "# It is assumed that for the \"predictor variable\" there should be no missing/ empty cells, and no \"out of bound\" values,\n",
    "# e.g. all the values in \"Price\" must be numeric, if there is any non-numeric value then error will be generated later.\n",
    "\n",
    "# The \"outcome\" variable is what we want to predict - Cold and Warm. In my analysis I relabelled all \"Hot\" as \"Warm\".\n",
    "# This was done as designing predictive model gets complex when the number of distinct output values are more than two.\n",
    "# Another challenges of working with records with \"Hot\" index for the given data is that there were only three records\n",
    "# with \"Hot\" as its Heat index - when the number of records with a given Heat index is very small in the historical \n",
    "# train data, then it does not lead to good prediction for the unknown data.\n",
    "\n",
    "# The data (specially numeric) should also be formatted on Excel, e.g. ensure that \"Price\" is a numeric quantity by \n",
    "# right clicking that column, click \"Format Cell\", select. And remove the $ sign and any comma.\n",
    "\n",
    "# Once this has been done, click on \"save as\" in excel, and then save the file as a csv file in the same folder.\n",
    "\n",
    "#flip_df = pd.read_csv('heat.csv')\n",
    "flip_df = pd.read_csv('Heat Exercise 5.csv')\n",
    "flip_df.columns = [s.strip().replace(' ', '_') for s in flip_df.columns]\n",
    "predictors = ['Family', 'Qty', 'Price_(Ea.)', 'EOL_time', 'Historic_Customers'] # 'Price_(Ea.)', 'EOL_time', 'Historic_Customers', \n",
    "outcome = 'Heat_Encode'\n",
    "#outcome = 'Heat'\n",
    "\n",
    "X1=flip_df[predictors]\n",
    "y1=flip_df[outcome]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Family</th>\n",
       "      <th>Historic_Customers</th>\n",
       "      <th>EOL</th>\n",
       "      <th>EOL_time</th>\n",
       "      <th>Finished_Goods_Inventory</th>\n",
       "      <th>Description</th>\n",
       "      <th>Qty</th>\n",
       "      <th>Price_(Ea.)</th>\n",
       "      <th>Total_Line_Value</th>\n",
       "      <th>Est_Die</th>\n",
       "      <th>Comments</th>\n",
       "      <th>Heat</th>\n",
       "      <th>Heat_Encode</th>\n",
       "      <th>Flip_Stock</th>\n",
       "      <th>Sales_History</th>\n",
       "      <th>Total_Quotes</th>\n",
       "      <th>Quoted_Amount</th>\n",
       "      <th>Authorized_Availability</th>\n",
       "      <th>Authorized_Price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>DSP56311</td>\n",
       "      <td>85</td>\n",
       "      <td>2018</td>\n",
       "      <td>2</td>\n",
       "      <td>DSP56311VF150R2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1967</td>\n",
       "      <td>34.76</td>\n",
       "      <td>$68,372.92</td>\n",
       "      <td>15250</td>\n",
       "      <td>All PN's listed here</td>\n",
       "      <td>Cold</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1,759</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>DSP56311</td>\n",
       "      <td>85</td>\n",
       "      <td>2018</td>\n",
       "      <td>2</td>\n",
       "      <td>DSP56311VL150</td>\n",
       "      <td>NaN</td>\n",
       "      <td>538</td>\n",
       "      <td>34.76</td>\n",
       "      <td>$18,700.88</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Warm</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5,450</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>C1K</td>\n",
       "      <td>47</td>\n",
       "      <td>2017</td>\n",
       "      <td>3</td>\n",
       "      <td>M83159G13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>99</td>\n",
       "      <td>13.41</td>\n",
       "      <td>$1,327.59</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Cold</td>\n",
       "      <td>0</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>C1K</td>\n",
       "      <td>47</td>\n",
       "      <td>2017</td>\n",
       "      <td>3</td>\n",
       "      <td>M83160G13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>314</td>\n",
       "      <td>11.95</td>\n",
       "      <td>$3,752.30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Cold</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>84</td>\n",
       "      <td>$14.80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>C1K</td>\n",
       "      <td>47</td>\n",
       "      <td>2017</td>\n",
       "      <td>3</td>\n",
       "      <td>M83240G13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>281</td>\n",
       "      <td>13.41</td>\n",
       "      <td>$3,768.21</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Cold</td>\n",
       "      <td>0</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",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Family  Historic_Customers   EOL  EOL_time Finished_Goods_Inventory  \\\n",
       "0  DSP56311                  85  2018         2          DSP56311VF150R2   \n",
       "1  DSP56311                  85  2018         2            DSP56311VL150   \n",
       "2       C1K                  47  2017         3                M83159G13   \n",
       "3       C1K                  47  2017         3                M83160G13   \n",
       "4       C1K                  47  2017         3                M83240G13   \n",
       "\n",
       "   Description   Qty  Price_(Ea.) Total_Line_Value Est_Die  \\\n",
       "0          NaN  1967        34.76       $68,372.92   15250   \n",
       "1          NaN   538        34.76       $18,700.88     NaN   \n",
       "2          NaN    99        13.41        $1,327.59     NaN   \n",
       "3          NaN   314        11.95        $3,752.30     NaN   \n",
       "4          NaN   281        13.41        $3,768.21     NaN   \n",
       "\n",
       "               Comments  Heat  Heat_Encode  Flip_Stock Sales_History  \\\n",
       "0  All PN's listed here  Cold            0         NaN           NaN   \n",
       "1                   NaN  Warm            1         NaN           NaN   \n",
       "2                   NaN  Cold            0         NaN           NaN   \n",
       "3                   NaN  Cold            0         NaN           NaN   \n",
       "4                   NaN  Cold            0         NaN           NaN   \n",
       "\n",
       "   Total_Quotes Quoted_Amount Authorized_Availability Authorized_Price  \n",
       "0           3.0         1,759                     NaN              NaN  \n",
       "1           2.0         5,450                     NaN              NaN  \n",
       "2           NaN           NaN                     NaN              NaN  \n",
       "3           NaN           NaN                      84           $14.80  \n",
       "4           NaN           NaN                     NaN              NaN  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# In this cell, the data which has been fetched from the csv file (xlsx -> csv) is displayed for visual inspection.\n",
    "# Only the first 5 rows are display by default, if you would like to see larger number of rows, enter the value \n",
    "# inside the paranthesis e.g. flip_df(7)\n",
    "\n",
    "flip_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The first line encodes the categorical data. The second line splits the data into training and validation data.\n",
    "\n",
    "# In this cell, we are only reading \"historical data\" with known actual value of the heat index, and the records\n",
    "# with unknown heat index value (labelled as X_new). The number of cell with known heat index value is listed within \n",
    "# the square brackets [0:72]. When the number of records changes, this value should also change.\n",
    "\n",
    "# The records with unknown heat index, and known heat index should be in the same csv data file. \n",
    "# Note here that Python labels the first record as zero, whereas in Excel the first record is labelled as one. In Excel,\n",
    "# the first row is used to label to columns.\n",
    "\n",
    "X1 = pd.get_dummies(X1, prefix_sep='_', drop_first=False)\n",
    "\n",
    "X = X1.loc[0:72]\n",
    "y = y1.loc[0:72]\n",
    "\n",
    "X_new = X1.loc[73:73]\n",
    "\n",
    "train_X, valid_X, train_y, valid_y = train_test_split(X, y, test_size=0.4, random_state=1)"
   ]
  },
  {
   "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>Qty</th>\n",
       "      <th>Price_(Ea.)</th>\n",
       "      <th>EOL_time</th>\n",
       "      <th>Historic_Customers</th>\n",
       "      <th>Family_C1K</th>\n",
       "      <th>Family_DSP56311</th>\n",
       "      <th>Family_Galaxy / MSC8122</th>\n",
       "      <th>Family_Goldfinger / MPC755</th>\n",
       "      <th>Family_MPC823 / T823</th>\n",
       "      <th>Family_MPC8241-D</th>\n",
       "      <th>Family_MPC8540 / PQ37</th>\n",
       "      <th>Family_MPC8541 / PQ37L</th>\n",
       "      <th>Family_MPC8555 / PQ37L</th>\n",
       "      <th>Family_MPC8560 / PQ37</th>\n",
       "      <th>Family_MPC857</th>\n",
       "      <th>Family_MPC862</th>\n",
       "      <th>Family_P1017</th>\n",
       "      <th>Family_P1023</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1967</td>\n",
       "      <td>34.76</td>\n",
       "      <td>2</td>\n",
       "      <td>85</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>538</td>\n",
       "      <td>34.76</td>\n",
       "      <td>2</td>\n",
       "      <td>85</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>99</td>\n",
       "      <td>13.41</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>314</td>\n",
       "      <td>11.95</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>281</td>\n",
       "      <td>13.41</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>68</td>\n",
       "      <td>5637</td>\n",
       "      <td>117.20</td>\n",
       "      <td>5</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>69</td>\n",
       "      <td>978</td>\n",
       "      <td>36.72</td>\n",
       "      <td>1</td>\n",
       "      <td>3</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>70</td>\n",
       "      <td>167</td>\n",
       "      <td>41.06</td>\n",
       "      <td>1</td>\n",
       "      <td>3</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>71</td>\n",
       "      <td>417</td>\n",
       "      <td>40.82</td>\n",
       "      <td>1</td>\n",
       "      <td>15</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>72</td>\n",
       "      <td>797</td>\n",
       "      <td>46.94</td>\n",
       "      <td>1</td>\n",
       "      <td>15</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",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>73 rows × 18 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     Qty  Price_(Ea.)  EOL_time  Historic_Customers  Family_C1K  \\\n",
       "0   1967        34.76         2                  85           0   \n",
       "1    538        34.76         2                  85           0   \n",
       "2     99        13.41         3                  47           1   \n",
       "3    314        11.95         3                  47           1   \n",
       "4    281        13.41         3                  47           1   \n",
       "..   ...          ...       ...                 ...         ...   \n",
       "68  5637       117.20         5                  16           0   \n",
       "69   978        36.72         1                   3           0   \n",
       "70   167        41.06         1                   3           0   \n",
       "71   417        40.82         1                  15           0   \n",
       "72   797        46.94         1                  15           0   \n",
       "\n",
       "    Family_DSP56311  Family_Galaxy / MSC8122  Family_Goldfinger / MPC755  \\\n",
       "0                 1                        0                           0   \n",
       "1                 1                        0                           0   \n",
       "2                 0                        0                           0   \n",
       "3                 0                        0                           0   \n",
       "4                 0                        0                           0   \n",
       "..              ...                      ...                         ...   \n",
       "68                0                        1                           0   \n",
       "69                0                        0                           0   \n",
       "70                0                        0                           0   \n",
       "71                0                        0                           0   \n",
       "72                0                        0                           0   \n",
       "\n",
       "    Family_MPC823 / T823  Family_MPC8241-D  Family_MPC8540 / PQ37  \\\n",
       "0                      0                 0                      0   \n",
       "1                      0                 0                      0   \n",
       "2                      0                 0                      0   \n",
       "3                      0                 0                      0   \n",
       "4                      0                 0                      0   \n",
       "..                   ...               ...                    ...   \n",
       "68                     0                 0                      0   \n",
       "69                     0                 0                      0   \n",
       "70                     0                 0                      0   \n",
       "71                     0                 0                      0   \n",
       "72                     0                 0                      0   \n",
       "\n",
       "    Family_MPC8541 / PQ37L  Family_MPC8555 / PQ37L  Family_MPC8560 / PQ37  \\\n",
       "0                        0                       0                      0   \n",
       "1                        0                       0                      0   \n",
       "2                        0                       0                      0   \n",
       "3                        0                       0                      0   \n",
       "4                        0                       0                      0   \n",
       "..                     ...                     ...                    ...   \n",
       "68                       0                       0                      0   \n",
       "69                       0                       0                      0   \n",
       "70                       0                       0                      0   \n",
       "71                       0                       0                      0   \n",
       "72                       0                       0                      0   \n",
       "\n",
       "    Family_MPC857  Family_MPC862  Family_P1017  Family_P1023  \n",
       "0               0              0             0             0  \n",
       "1               0              0             0             0  \n",
       "2               0              0             0             0  \n",
       "3               0              0             0             0  \n",
       "4               0              0             0             0  \n",
       "..            ...            ...           ...           ...  \n",
       "68              0              0             0             0  \n",
       "69              0              0             1             0  \n",
       "70              0              0             1             0  \n",
       "71              0              0             0             1  \n",
       "72              0              0             0             1  \n",
       "\n",
       "[73 rows x 18 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# In this I can visually inspect the record with unknown het index. Categorical data has been encoded.\n",
    "\n",
    "X1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "********************\n",
    "<b><u>Classifiers</u></b><br>\n",
    "Now that I have fetched the clean data, I can now start training-testing our data with various classification algorithm.\n",
    "I am using the ensemble learning approach to make prediction - Essentially this approach makes use of \"Majority Voting\"\n",
    "approach to make decision, i.e. I train more than one classifier, and then when all the classifiers have been trained,\n",
    "then for the unknown data these classifiers \"vote\" whether the outcome will be \"Cold\" or \"Warm\". Decision is then made in favour of heat index value which gets the highest number of votes.\n",
    "\n",
    "However the performance of the \"Majority Voting\" approach is dependent on choosing suitable classifiers which can participate\n",
    "in the voting decision. Theoretically, a classifier which achieves an accuracy of >50% is selected, however in practise the\n",
    "threshold value is user-defined.\n",
    "********************"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from dmba import plotDecisionTree\n",
    "fullClassTree = DecisionTreeClassifier(random_state=1)\n",
    "fullClassTree.fit(train_X, train_y)\n",
    "\n",
    "plotDecisionTree(fullClassTree, feature_names=train_X.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Improved score:  0.586111111111111\n",
      "Improved parameters:  {'max_depth': 5, 'min_impurity_decrease': 0.0009, 'min_samples_split': 10}\n"
     ]
    }
   ],
   "source": [
    "# This is our first classifier (known as Decision Tree). No changes need to made here.\n",
    "\n",
    "# Start with an initial guess for parameters\n",
    "param_grid = {\n",
    "'max_depth': [10, 20, 30, 40],\n",
    "'min_samples_split': [20, 40, 60, 80, 100],\n",
    "'min_impurity_decrease': [0, 0.0005, 0.001, 0.005, 0.01],\n",
    "}\n",
    "\n",
    "gridSearch = GridSearchCV(DecisionTreeClassifier(random_state=1), param_grid, cv=5, n_jobs=-1) # n_jobs=-1 will utilize all available CPUs\n",
    "gridSearch.fit(train_X, train_y)\n",
    "\n",
    "#print('Initial score: ', gridSearch.best_score_)\n",
    "#print('Initial parameters: ', gridSearch.best_params_)\n",
    "\n",
    "# Adapt grid based on result from initial grid search\n",
    "param_grid = {\n",
    "'max_depth': list(range(2, 16)), # 14 values\n",
    "'min_samples_split': list(range(10, 22)), # 11 values\n",
    "'min_impurity_decrease': [0.0009, 0.001, 0.0011], # 3 values\n",
    "}\n",
    "gridSearch = GridSearchCV(DecisionTreeClassifier(random_state=1), param_grid, cv=5,\n",
    "n_jobs=-1)\n",
    "gridSearch.fit(train_X, train_y)\n",
    "print('Improved score: ', gridSearch.best_score_)\n",
    "print('Improved parameters: ', gridSearch.best_params_)\n",
    "bestClassTree = gridSearch.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.7333)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    5    7\n",
      "  Warm    1   17\n"
     ]
    }
   ],
   "source": [
    "# This cell will output the accuracy of the Decision Tree. As the accuracy is 73% ( more than 50%) for the validation \n",
    "# data, I can theoretically admit this classifier in the Majority Voting later.\n",
    "classificationSummary(valid_y, gridSearch.predict(valid_X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.7667)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    8    4\n",
      "  Warm    3   15\n"
     ]
    }
   ],
   "source": [
    "# This cell trains a second classifier (Random Forest) and outputs the accuracy. No changes need to be made in this cell.\n",
    "\n",
    "rf = RandomForestClassifier(n_estimators=500, random_state=1)\n",
    "rf.fit(train_X, train_y)\n",
    "classificationSummary(valid_y, rf.predict(valid_X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.7667)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    8    4\n",
      "  Warm    3   15\n"
     ]
    }
   ],
   "source": [
    "# This cell trains a third classifier (Gradient Boosting) and outputs the accuracy. No changes need to be made in this cell.\n",
    "\n",
    "boost = GradientBoostingClassifier()\n",
    "boost.fit(train_X, train_y)\n",
    "classificationSummary(valid_y, boost.predict(valid_X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.6333)\n",
      "\n",
      "       Prediction\n",
      "Actual  0  1\n",
      "     0  8  4\n",
      "     1  7 11\n"
     ]
    }
   ],
   "source": [
    "# This cell trains a fourth classifier (Neural Net) and outputs the accuracy. No changes need to be made in this cell.\n",
    "# The cell may output a message such \"STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\" However this should not be a cause\n",
    "# of any concern. No changes need to be made here.\n",
    "\n",
    "classes = sorted(y.unique())\n",
    "clf = MLPClassifier(hidden_layer_sizes=(500), activation='logistic', solver='lbfgs', random_state=1, max_iter=20000)\n",
    "clf.fit(train_X, train_y)\n",
    "classificationSummary(valid_y, clf.predict(valid_X), class_names=classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.7333)\n",
      "\n",
      "       Prediction\n",
      "Actual  0  1\n",
      "     0  8  4\n",
      "     1  4 14\n"
     ]
    }
   ],
   "source": [
    "# This cell trains a fifth classifier (Neural Net with different parameters) and outputs the accuracy. \n",
    "# No changes need to be made in this cell.\n",
    "\n",
    "clf2 = MLPClassifier(hidden_layer_sizes=(200), activation='tanh', solver='lbfgs', random_state=1, max_iter=20000)\n",
    "clf2.fit(train_X, train_y)\n",
    "#clf2.predict(X)\n",
    "classificationSummary(valid_y, clf2.predict(valid_X), class_names=classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.6333)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    7    5\n",
      "  Warm    6   12\n"
     ]
    }
   ],
   "source": [
    "# This cell trains a sixth classifier (Logistic Regression) and outputs the accuracy. \n",
    "# No changes need to be made in this cell.\n",
    "\n",
    "logit_reg = LogisticRegression(penalty=\"l2\", C=1e42, solver='liblinear')\n",
    "logit_reg.fit(train_X, train_y)\n",
    "classificationSummary(valid_y, logit_reg.predict(valid_X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.7333)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    6    6\n",
      "  Warm    2   16\n"
     ]
    }
   ],
   "source": [
    "# This cell trains a seventh classifier (SVM with polynomial of degree 5) and outputs the accuracy. \n",
    "# No changes need to be made in this cell.\n",
    "\n",
    "h1 = svm.SVC(kernel='poly', degree=5, gamma=0.001, C=1.0) # Polynomial degree 5 SVC #probability=True\n",
    "h1.fit(train_X, train_y)\n",
    "classificationSummary(valid_y, h1.predict(valid_X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.6667)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    7    5\n",
      "  Warm    5   13\n"
     ]
    }
   ],
   "source": [
    "# This cell trains a eighth classifier (SVM with polynomial of degree 2) and outputs the accuracy. \n",
    "# No changes need to be made in this cell.\n",
    "\n",
    "h2 = svm.SVC(kernel='poly', degree=2, gamma=0.001, C=1.0) # probability=True\n",
    "h2.fit(train_X, train_y)\n",
    "classificationSummary(valid_y, h2.predict(valid_X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix (Accuracy 0.6667)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    9    3\n",
      "  Warm    7   11\n"
     ]
    }
   ],
   "source": [
    "# This cell trains an ninth classifier (KNN with two classes) and outputs the accuracy. \n",
    "# No changes need to be made in this cell.\n",
    "\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "knn = KNeighborsClassifier(2)\n",
    "knn.fit(train_X, train_y)\n",
    "classificationSummary(valid_y, knn.predict(valid_X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***************\n",
    "<u><b>Classifiers Selection</b></u><br>\n",
    "Now that I have trained various classifiers, I will now select those classifers with accuracy >50%, and then run my final\n",
    "evaluation, and use it to predict the unknown data values. For the voting decision not to result in a tie, an odd number\n",
    "of classifier should be selected.\n",
    "<br>\n",
    "<br>\n",
    "Comment out those classifier using # which are not needed because of their relatively poor accuracy performance. I am selecting the top 5 classifiers for my model.\n",
    "***************"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# MAJORITY VOTING\n",
    "# Now that various classifiers have been trained use majority voting here.\n",
    "# This cell can take FEW minutes to execute depending on the processor of the computer system used.\n",
    "\n",
    "eclf1 = VotingClassifier(estimators=[\n",
    "    \n",
    "                                     ('GS', gridSearch),\n",
    "                                     ('RF', rf),  \n",
    "                                     ('Boost', boost),\n",
    "                                   #  ('NN1', clf), \n",
    "                                     ('NN2', clf2),\n",
    "                                    # ('LOG', logit_reg),\n",
    "                                     ('SVM1', h1),\n",
    "                                    # ('SVM2', h2),\n",
    "                                    # ('KNN', knn)\n",
    "    \n",
    "                                    ], voting='hard')\n",
    "\n",
    "eclf1 = eclf1.fit(train_X, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of validation data: \n",
      "Confusion Matrix (Accuracy 0.8000)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold    8    4\n",
      "  Warm    2   16\n",
      "Accuracy of full data: \n",
      "Confusion Matrix (Accuracy 0.9178)\n",
      "\n",
      "       Prediction\n",
      "Actual Cold Warm\n",
      "  Cold   30    4\n",
      "  Warm    2   37\n"
     ]
    }
   ],
   "source": [
    "print('Accuracy of validation data: ')\n",
    "classificationSummary(valid_y, eclf1.predict(valid_X), class_names=['Cold', 'Warm'])\n",
    "\n",
    "print('Accuracy of full data: ')\n",
    "classificationSummary(y, eclf1.predict(X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# In this cell I generate \"belief probability\" of the decision. The choice of classifier which I select is based on\n",
    "# their accuracy, and may not necessarily reflect classifiers selected during the MAJORITY VOTING cell.\n",
    "# In the last line, only those prob values should be included in the average_proba calculatin which were not \n",
    "# commented. The number with which it is divided should reflect the number of variables which are being added up.\n",
    "\n",
    "# These steps are needed if the analyst wishes to manually validate heat index decision for records with relatively lower\n",
    "# probabilities.\n",
    "\n",
    "#gridSearch_proba = gridSearch.predict_proba(X)\n",
    "#rf_proba = rf.predict_proba(X)\n",
    "#boost_proba = boost.predict_proba(X)\n",
    "#clf_proba = clf.predict_proba(X)\n",
    "#clf_proba2 = clf2.predict_proba(X_new)\n",
    "#logit_reg_proba = logit_reg.predict_proba(X_new)\n",
    "#h1_proba = h1.predict_proba(X_new) # can not be uncommented, as there is no probability associated with SVM !!!\n",
    "#h2_proba = h2.predict_proba(X_new) # can not be uncommented, as there is no probability associated with SVM !!!\n",
    "#knn_proba = knn.predict_proba(X_new)\n",
    "\n",
    "#average_proba=(gridSearch_proba+rf_proba+boost_proba+clf_proba)/4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#final_predict=eclf1.predict(X)\n",
    "\n",
    "#final_result = pd.DataFrame({'actual': y,\n",
    "#'cold prob.': [p[0] for p in average_proba],\n",
    "#'warm prob.': [p[1] for p in average_proba],\n",
    "#'predicted': final_predict })\n",
    "\n",
    "#print(final_result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The result (cold/ warm prob. and predicted) produced in the previous cell will now be imported to the file \"result.csv\". \n",
    "# The analyst can aggregate this result with the result in the original fule \"heat.csv\" to perform any further function\n",
    "# as required.\n",
    "\n",
    "#final_result.to_csv(r'C:\\Users\\jalaluddin\\Desktop\\GBC B412\\Data Project Capstone BUS 4045\\Assignments\\imbalance\\result.csv', index = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Classifiers which we choose for majority voting\n",
    "# We plot the AUC of the those classifiers which have been selected for the Majority Voting. \n",
    "\n",
    "y_pred_proba_gridSearch = gridSearch.predict_proba(valid_X)[::,1]\n",
    "y_pred_proba_rf = rf.predict_proba(valid_X)[::,1]\n",
    "y_pred_proba_boost = boost.predict_proba(valid_X)[::,1]\n",
    "y_pred_proba_clf2 = clf2.predict_proba(valid_X)[::,1]\n",
    "y_pred_proba_knn = knn.predict_proba(valid_X)[::,1]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "fpr1, tpr1, _ = metrics.roc_curve(valid_y,  y_pred_proba_rf)\n",
    "fpr2, tpr2, _ = metrics.roc_curve(valid_y,  y_pred_proba_boost)\n",
    "fpr3, tpr3, _ = metrics.roc_curve(valid_y,  y_pred_proba_gridSearch)\n",
    "fpr4, tpr4, _ = metrics.roc_curve(valid_y,  y_pred_proba_clf2)\n",
    "fpr5, tpr5, _ = metrics.roc_curve(valid_y,  y_pred_proba_knn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#https://scikit-learn.org/stable/auto_examples/model_selection/plot_roc.html\n",
    "# https://stackoverflow.com/questions/25009284/how-to-plot-roc-curve-in-python\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "auc1 = metrics.roc_auc_score(valid_y, y_pred_proba_rf)\n",
    "auc2 = metrics.roc_auc_score(valid_y, y_pred_proba_boost)\n",
    "auc3 = metrics.roc_auc_score(valid_y, y_pred_proba_gridSearch)\n",
    "auc4 = metrics.roc_auc_score(valid_y, y_pred_proba_clf2)\n",
    "auc5 = metrics.roc_auc_score(valid_y, y_pred_proba_knn)\n",
    "\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(fpr1,tpr1, lw=2, label='Random Forest ROC curve (area = %0.2f)' % auc1) #color='navy', \n",
    "plt.plot(fpr2,tpr2, lw=2, label='Gradient Boost ROC curve (area = %0.2f)' % auc2) #color='navy', \n",
    "plt.plot(fpr3,tpr3, lw=2, label='Decision Tree ROC curve (area = %0.2f)' % auc3) #color='navy', \n",
    "plt.plot(fpr4,tpr4, lw=2, label='Neural Network ROC curve (area = %0.2f)' % auc4) #color='navy', \n",
    "plt.plot(fpr5,tpr5, lw=2, label='KNN ROC curve (area = %0.2f)' % auc5) #color='navy', \n",
    "\n",
    "plt.plot([0, 1], [0, 1], lw=2, linestyle='--')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.savefig(\"classifiers1.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# MAJORITY VOTING\n",
    "# Soft voting so that a probability estimate can be obtained fr AUC plotting.\n",
    "\n",
    "eclf2 = VotingClassifier(estimators=[\n",
    "    \n",
    "                                     ('GS', gridSearch),\n",
    "                                     ('RF', rf),  \n",
    "                                     ('Boost', boost),\n",
    "                                    # ('NN1', clf), \n",
    "                                     ('NN2', clf2),\n",
    "                                    # ('LOG', logit_reg),\n",
    "                                    # ('SVM1', h1),\n",
    "                                    # ('SVM2', h2),\n",
    "                                     ('KNN', knn)\n",
    "    \n",
    "                                    ], voting='soft')\n",
    "\n",
    "eclf2 = eclf2.fit(train_X, train_y)\n",
    "\n",
    "print('Accuracy of validation data: ')\n",
    "classificationSummary(valid_y, eclf2.predict(valid_X), class_names=['Cold', 'Warm'])\n",
    "\n",
    "print('Accuracy of full data: ')\n",
    "classificationSummary(y, eclf2.predict(X), class_names=['Cold', 'Warm'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Plotting the AUC for the Soft Voting AUC graph.\n",
    "\n",
    "y_pred_proba_eclf2 = eclf2.predict_proba(valid_X)[::,1]\n",
    "fpr6, tpr6, _ = metrics.roc_curve(valid_y,  y_pred_proba_eclf2)\n",
    "auc6 = metrics.roc_auc_score(valid_y, y_pred_proba_eclf2)\n",
    "plt.figure()\n",
    "plt.plot(fpr6,tpr6, lw=2, label='Majority Voting ROC curve (area = %0.2f)' % auc6) #color='navy', \n",
    "plt.plot([0, 1], [0, 1], lw=2, linestyle='--')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.savefig(\"majority_class.png\")"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}