BM Full.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Big mart sales with full intuition\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "We will explore the problem in following stages:\n",
    "\n",
    "Hypothesis Generation – understanding the problem better by brainstorming possible factors that can impact the outcome\n",
    "\n",
    "Data Exploration – looking at categorical and continuous feature summaries and making inferences about the data.\n",
    "\n",
    "Data Cleaning – imputing missing values in the data and checking for outliers.\n",
    "\n",
    "Feature Engineering – modifying existing variables and creating new ones for analysis.\n",
    "\n",
    "Model Building – making predictive models on the data\n",
    "\n",
    "First we will load required libraries and the data and check each and every column and understand what effect it can have. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#importing the required libraries\n",
    "import pandas as pd \n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:8: FutureWarning: Sorting because non-concatenation axis is not aligned. A future version\n",
      "of pandas will change to not sort by default.\n",
      "\n",
      "To accept the future behavior, pass 'sort=False'.\n",
      "\n",
      "To retain the current behavior and silence the warning, pass 'sort=True'.\n",
      "\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "# importing the data\n",
    "train= pd.read_csv(\"Train_UWu5bXk.csv\")\n",
    "test=  pd.read_csv(\"Test_u94Q5KV.csv\")\n",
    "# merging train and test together so that all parameters are scaled together \n",
    "#created source because we have to separate them afterwards.\n",
    "train['source']='train'\n",
    "test['source']='test'\n",
    "data= pd.concat([train,test],axis=0,ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(14204, 13) (8523, 13) (5681, 12)\n"
     ]
    }
   ],
   "source": [
    "print(data.shape,train.shape,test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Item_Fat_Content                0\n",
       "Item_Identifier                 0\n",
       "Item_MRP                        0\n",
       "Item_Outlet_Sales            5681\n",
       "Item_Type                       0\n",
       "Item_Visibility                 0\n",
       "Item_Weight                  2439\n",
       "Outlet_Establishment_Year       0\n",
       "Outlet_Identifier               0\n",
       "Outlet_Location_Type            0\n",
       "Outlet_Size                  4016\n",
       "Outlet_Type                     0\n",
       "source                          0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Now we are checking if any columns have null vale and how much\n",
    "#here outlet_sales can be ignored as it is for missing in test data\n",
    "data.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "here we have missing values in weight and outlet size, which we will impute in our next steps.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Determine the average weight per item_identifier:\n",
    "\n",
    "item_avg_weight = data.pivot_table(values='Item_Weight', index='Item_Identifier', aggfunc='mean')\n",
    "# item_avg_weight.loc['FDP10']\n",
    "\n",
    "# Mean_values_Item_Type_data=data.groupby('Item_Identifier')['Item_Weight'].mean()\n",
    "# Mean_values_Item_Type_data\n",
    "\n",
    "#Taking the missing values\n",
    "misval= data['Item_Weight'].isnull()\n",
    "# data.loc[data['Item_Identifier']=='FDP10','Item_Weight']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Here we will replace missing value with there respective Item_Identifier mean value\n",
    "# df.loc[rows,columns]\n",
    "\n",
    "data.loc[misval,'Item_Weight']= data.loc[misval,'Item_Identifier'].apply(lambda x: item_avg_weight.loc[x])\n",
    "# data.loc[miss_bool,'Item_Weight'] = data.loc[miss_bool,'Item_Identifier'].apply(lambda x: item_avg_weight.loc[x])\n",
    "\n",
    "data['Item_Weight'].isnull().sum()"
   ]
  },
  {
   "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 th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>Outlet_Type</th>\n",
       "      <th>Grocery Store</th>\n",
       "      <th>Supermarket Type1</th>\n",
       "      <th>Supermarket Type2</th>\n",
       "      <th>Supermarket Type3</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Outlet_Size</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>High</th>\n",
       "      <td>0</td>\n",
       "      <td>1553</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Medium</th>\n",
       "      <td>0</td>\n",
       "      <td>1550</td>\n",
       "      <td>1546</td>\n",
       "      <td>1559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Small</th>\n",
       "      <td>880</td>\n",
       "      <td>3100</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Outlet_Type  Grocery Store  Supermarket Type1  Supermarket Type2  \\\n",
       "Outlet_Size                                                        \n",
       "High                     0               1553                  0   \n",
       "Medium                   0               1550               1546   \n",
       "Small                  880               3100                  0   \n",
       "\n",
       "Outlet_Type  Supermarket Type3  \n",
       "Outlet_Size                     \n",
       "High                         0  \n",
       "Medium                    1559  \n",
       "Small                        0  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Imputing the Outlet_Size missing values\n",
    "#filling value of Outlet_Size.\n",
    "#Hypothesis: Outlet_Size depends on Oulet Type and Outlet__Location_Type\n",
    "\n",
    "twowaytable1=pd.crosstab(data['Outlet_Size'],data['Outlet_Type'])\n",
    "twowaytable1\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Outlet_Location_Type</th>\n",
       "      <th>Tier 1</th>\n",
       "      <th>Tier 2</th>\n",
       "      <th>Tier 3</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Outlet_Size</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>High</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Medium</th>\n",
       "      <td>1550</td>\n",
       "      <td>0</td>\n",
       "      <td>3105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Small</th>\n",
       "      <td>2430</td>\n",
       "      <td>1550</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Outlet_Location_Type  Tier 1  Tier 2  Tier 3\n",
       "Outlet_Size                                 \n",
       "High                       0       0    1553\n",
       "Medium                  1550       0    3105\n",
       "Small                   2430    1550       0"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "twowaytable2=pd.crosstab(data['Outlet_Size'],data['Outlet_Location_Type'])\n",
    "twowaytable2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False    14204\n",
       "Name: Outlet_Type, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# now by both tab we can conclude, grocery store and tier 2 city has small size\n",
    "# so we will map them.\n",
    "d={'Grocery Store':'Small', 'Tier 2': 'Small',\n",
    "  'Supermarket Type2':'Medium', 'Supermarket Type3': 'Medium'}\n",
    "s= data['Outlet_Type'].map(d)\n",
    "a= data['Outlet_Location_Type'].map(d)\n",
    "#map= Used for substituting each value in a Series with another value, that may be derived from a function, \n",
    "# a dict or a Series.\n",
    "data.Outlet_Size=data.Outlet_Size.combine_first(s)\n",
    "#combine_first: Used for Combine Series values, choosing the calling Series’s values first.\n",
    "data.Outlet_Size=data.Outlet_Size.combine_first(a)\n",
    "data['Outlet_Type'].isnull().value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>Item_MRP</th>\n",
       "      <th>Item_Outlet_Sales</th>\n",
       "      <th>Item_Visibility</th>\n",
       "      <th>Item_Weight</th>\n",
       "      <th>Outlet_Establishment_Year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>14204.000000</td>\n",
       "      <td>8523.000000</td>\n",
       "      <td>14204.000000</td>\n",
       "      <td>14204.000000</td>\n",
       "      <td>14204.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>141.004977</td>\n",
       "      <td>2181.288914</td>\n",
       "      <td>0.065953</td>\n",
       "      <td>12.793380</td>\n",
       "      <td>1997.830681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>62.086938</td>\n",
       "      <td>1706.499616</td>\n",
       "      <td>0.051459</td>\n",
       "      <td>4.651716</td>\n",
       "      <td>8.371664</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>31.290000</td>\n",
       "      <td>33.290000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.555000</td>\n",
       "      <td>1985.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>94.012000</td>\n",
       "      <td>834.247400</td>\n",
       "      <td>0.027036</td>\n",
       "      <td>8.710000</td>\n",
       "      <td>1987.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>142.247000</td>\n",
       "      <td>1794.331000</td>\n",
       "      <td>0.054021</td>\n",
       "      <td>12.600000</td>\n",
       "      <td>1999.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>185.855600</td>\n",
       "      <td>3101.296400</td>\n",
       "      <td>0.094037</td>\n",
       "      <td>16.750000</td>\n",
       "      <td>2004.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>266.888400</td>\n",
       "      <td>13086.964800</td>\n",
       "      <td>0.328391</td>\n",
       "      <td>21.350000</td>\n",
       "      <td>2009.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Item_MRP  Item_Outlet_Sales  Item_Visibility   Item_Weight  \\\n",
       "count  14204.000000        8523.000000     14204.000000  14204.000000   \n",
       "mean     141.004977        2181.288914         0.065953     12.793380   \n",
       "std       62.086938        1706.499616         0.051459      4.651716   \n",
       "min       31.290000          33.290000         0.000000      4.555000   \n",
       "25%       94.012000         834.247400         0.027036      8.710000   \n",
       "50%      142.247000        1794.331000         0.054021     12.600000   \n",
       "75%      185.855600        3101.296400         0.094037     16.750000   \n",
       "max      266.888400       13086.964800         0.328391     21.350000   \n",
       "\n",
       "       Outlet_Establishment_Year  \n",
       "count               14204.000000  \n",
       "mean                 1997.830681  \n",
       "std                     8.371664  \n",
       "min                  1985.000000  \n",
       "25%                  1987.000000  \n",
       "50%                  1999.000000  \n",
       "75%                  2004.000000  \n",
       "max                  2009.000000  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#now we will check the basic properties of numerical value in our data\n",
    "data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we judged that item visibility has minimum value of 0, which is strange as if item is present then it should be visible.\n",
    "and establishment year ranges from 1985 to 2009, but year wont help instead how old it it will help us as majot factor in our prediction.\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#replacing 0 visibility with mean visibility\n",
    "# data['Item_Visibility'].replace(0.0,np.nan)\n",
    "# data['Item_Visibility'].fillna(data['Item_Visibility'].mean(), inplace=True)\n",
    "# data['Item_Visibility'].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of 0 values initially: 879\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([False])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Determine average visibility of a product\n",
    "visibility_avg = data.pivot_table(values='Item_Visibility', index='Item_Identifier', aggfunc='mean')\n",
    "\n",
    "#Impute 0 values with mean visibility of that product:\n",
    "miss_bool = (data['Item_Visibility'] == 0)\n",
    "print ('Number of 0 values initially: %d'%(sum(miss_bool)))\n",
    "\n",
    "data.loc[miss_bool,'Item_Visibility'] = data.loc[miss_bool,'Item_Identifier'].apply(lambda x: visibility_avg.loc[x])\n",
    "\n",
    "(data['Item_Visibility'] == 0).unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "the Item_Type variable has 16 categories which might prove to be very useful in analysis. So its a good idea to combine them. One way could be to manually assign a new category to each. But there’s a catch here. If you look at the Item_Identifier, i.e. the unique ID of each item, it starts with either FD, DR or NC. If you see the categories, these look like being Food, Drinks and Non-Consumables. So I’ve used the Item_Identifier variable to create a new column:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Food              10201\n",
       "Non-Consumable     2686\n",
       "Drinks             1317\n",
       "Name: Item_Type_Combined, dtype: int64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Get the first two characters of ID:\n",
    "data['Item_Type_Combined'] = data['Item_Identifier'].apply(lambda x: x[0:2])\n",
    "#Rename them to more intuitive categories:\n",
    "data['Item_Type_Combined'] = data['Item_Type_Combined'].map({'FD':'Food',\n",
    "                                                             'NC':'Non-Consumable',\n",
    "                                                             'DR':'Drinks'})\n",
    "data['Item_Type_Combined'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    14204.000000\n",
       "mean        21.169319\n",
       "std          8.371664\n",
       "min         10.000000\n",
       "25%         15.000000\n",
       "50%         20.000000\n",
       "75%         32.000000\n",
       "max         34.000000\n",
       "Name: Outlet_Years, dtype: float64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#getting no establishment year\n",
    "#Years:\n",
    "data['Outlet_Years'] = 2019 - data['Outlet_Establishment_Year']\n",
    "data['Outlet_Years'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Item_Fat_Content                 5\n",
       "Item_Identifier               1559\n",
       "Item_MRP                      8052\n",
       "Item_Outlet_Sales             3494\n",
       "Item_Type                       16\n",
       "Item_Visibility              13688\n",
       "Item_Weight                    628\n",
       "Outlet_Establishment_Year        9\n",
       "Outlet_Identifier               10\n",
       "Outlet_Location_Type             3\n",
       "Outlet_Size                      3\n",
       "Outlet_Type                      4\n",
       "source                           2\n",
       "Item_Type_Combined               3\n",
       "Outlet_Years                     9\n",
       "dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# now we will check for categorical variables, and how it has difference\n",
    "\n",
    "data.apply(lambda x: len(x.unique()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Item_Fat_Content              object\n",
       "Item_Identifier               object\n",
       "Item_MRP                     float64\n",
       "Item_Outlet_Sales            float64\n",
       "Item_Type                     object\n",
       "Item_Visibility              float64\n",
       "Item_Weight                  float64\n",
       "Outlet_Establishment_Year      int64\n",
       "Outlet_Identifier             object\n",
       "Outlet_Location_Type          object\n",
       "Outlet_Size                   object\n",
       "Outlet_Type                   object\n",
       "source                        object\n",
       "Item_Type_Combined            object\n",
       "Outlet_Years                   int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Item_Fat_Content',\n",
       " 'Item_Identifier',\n",
       " 'Item_Type',\n",
       " 'Outlet_Identifier',\n",
       " 'Outlet_Location_Type',\n",
       " 'Outlet_Size',\n",
       " 'Outlet_Type',\n",
       " 'source',\n",
       " 'Item_Type_Combined']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical_columns = [x for x in data.dtypes.index if data.dtypes[x]=='object']\n",
    "categorical_columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Item_Fat_Content\n",
      "Low Fat    8485\n",
      "Regular    4824\n",
      "LF          522\n",
      "reg         195\n",
      "low fat     178\n",
      "Name: Item_Fat_Content, dtype: int64\n",
      "False    14204\n",
      "Name: Item_Fat_Content, dtype: int64\n",
      "Item_Type\n",
      "Fruits and Vegetables    2013\n",
      "Snack Foods              1989\n",
      "Household                1548\n",
      "Frozen Foods             1426\n",
      "Dairy                    1136\n",
      "Baking Goods             1086\n",
      "Canned                   1084\n",
      "Health and Hygiene        858\n",
      "Meat                      736\n",
      "Soft Drinks               726\n",
      "Breads                    416\n",
      "Hard Drinks               362\n",
      "Others                    280\n",
      "Starchy Foods             269\n",
      "Breakfast                 186\n",
      "Seafood                    89\n",
      "Name: Item_Type, dtype: int64\n",
      "False    14204\n",
      "Name: Item_Type, dtype: int64\n",
      "Outlet_Location_Type\n",
      "Tier 3    5583\n",
      "Tier 2    4641\n",
      "Tier 1    3980\n",
      "Name: Outlet_Location_Type, dtype: int64\n",
      "False    14204\n",
      "Name: Outlet_Location_Type, dtype: int64\n",
      "Outlet_Size\n",
      "Small     7996\n",
      "Medium    4655\n",
      "High      1553\n",
      "Name: Outlet_Size, dtype: int64\n",
      "False    14204\n",
      "Name: Outlet_Size, dtype: int64\n",
      "Outlet_Type\n",
      "Supermarket Type1    9294\n",
      "Grocery Store        1805\n",
      "Supermarket Type3    1559\n",
      "Supermarket Type2    1546\n",
      "Name: Outlet_Type, dtype: int64\n",
      "False    14204\n",
      "Name: Outlet_Type, dtype: int64\n",
      "Item_Type_Combined\n",
      "Food              10201\n",
      "Non-Consumable     2686\n",
      "Drinks             1317\n",
      "Name: Item_Type_Combined, dtype: int64\n",
      "False    14204\n",
      "Name: Item_Type_Combined, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "for i in categorical_columns:\n",
    "    if i not in ['Item_Identifier','source','Outlet_Identifier']:\n",
    "        print(i)\n",
    "        print(data[i].value_counts())\n",
    "        print(data[i].isnull().value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Item_Type: Not all categories have substantial numbers. It looks like combining them can give better results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Low Fat    9185\n",
       "Regular    5019\n",
       "Name: Item_Fat_Content, dtype: int64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Change categories of low fat:\n",
    "\n",
    "data['Item_Fat_Content'].value_counts()\n",
    "\n",
    "\n",
    "data['Item_Fat_Content'] = data['Item_Fat_Content'].replace({'LF':'Low Fat',\n",
    "                                                             'reg':'Regular',\n",
    "                                                             'low fat':'Low Fat'})\n",
    "data['Item_Fat_Content'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>Item_Fat_Content</th>\n",
       "      <th>Item_Identifier</th>\n",
       "      <th>Item_MRP</th>\n",
       "      <th>Item_Outlet_Sales</th>\n",
       "      <th>Item_Type</th>\n",
       "      <th>Item_Visibility</th>\n",
       "      <th>Item_Weight</th>\n",
       "      <th>Outlet_Establishment_Year</th>\n",
       "      <th>Outlet_Identifier</th>\n",
       "      <th>Outlet_Location_Type</th>\n",
       "      <th>Outlet_Size</th>\n",
       "      <th>Outlet_Type</th>\n",
       "      <th>source</th>\n",
       "      <th>Item_Type_Combined</th>\n",
       "      <th>Outlet_Years</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Low Fat</td>\n",
       "      <td>FDA15</td>\n",
       "      <td>249.8092</td>\n",
       "      <td>3735.1380</td>\n",
       "      <td>Dairy</td>\n",
       "      <td>0.016047</td>\n",
       "      <td>9.30</td>\n",
       "      <td>1999</td>\n",
       "      <td>OUT049</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>Medium</td>\n",
       "      <td>Supermarket Type1</td>\n",
       "      <td>train</td>\n",
       "      <td>Food</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Regular</td>\n",
       "      <td>DRC01</td>\n",
       "      <td>48.2692</td>\n",
       "      <td>443.4228</td>\n",
       "      <td>Soft Drinks</td>\n",
       "      <td>0.019278</td>\n",
       "      <td>5.92</td>\n",
       "      <td>2009</td>\n",
       "      <td>OUT018</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>Medium</td>\n",
       "      <td>Supermarket Type2</td>\n",
       "      <td>train</td>\n",
       "      <td>Drinks</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Low Fat</td>\n",
       "      <td>FDN15</td>\n",
       "      <td>141.6180</td>\n",
       "      <td>2097.2700</td>\n",
       "      <td>Meat</td>\n",
       "      <td>0.016760</td>\n",
       "      <td>17.50</td>\n",
       "      <td>1999</td>\n",
       "      <td>OUT049</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>Medium</td>\n",
       "      <td>Supermarket Type1</td>\n",
       "      <td>train</td>\n",
       "      <td>Food</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Regular</td>\n",
       "      <td>FDX07</td>\n",
       "      <td>182.0950</td>\n",
       "      <td>732.3800</td>\n",
       "      <td>Fruits and Vegetables</td>\n",
       "      <td>0.017834</td>\n",
       "      <td>19.20</td>\n",
       "      <td>1998</td>\n",
       "      <td>OUT010</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>Small</td>\n",
       "      <td>Grocery Store</td>\n",
       "      <td>train</td>\n",
       "      <td>Food</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Low Fat</td>\n",
       "      <td>NCD19</td>\n",
       "      <td>53.8614</td>\n",
       "      <td>994.7052</td>\n",
       "      <td>Household</td>\n",
       "      <td>0.009780</td>\n",
       "      <td>8.93</td>\n",
       "      <td>1987</td>\n",
       "      <td>OUT013</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>High</td>\n",
       "      <td>Supermarket Type1</td>\n",
       "      <td>train</td>\n",
       "      <td>Non-Consumable</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Item_Fat_Content Item_Identifier  Item_MRP  Item_Outlet_Sales  \\\n",
       "0          Low Fat           FDA15  249.8092          3735.1380   \n",
       "1          Regular           DRC01   48.2692           443.4228   \n",
       "2          Low Fat           FDN15  141.6180          2097.2700   \n",
       "3          Regular           FDX07  182.0950           732.3800   \n",
       "4          Low Fat           NCD19   53.8614           994.7052   \n",
       "\n",
       "               Item_Type  Item_Visibility  Item_Weight  \\\n",
       "0                  Dairy         0.016047         9.30   \n",
       "1            Soft Drinks         0.019278         5.92   \n",
       "2                   Meat         0.016760        17.50   \n",
       "3  Fruits and Vegetables         0.017834        19.20   \n",
       "4              Household         0.009780         8.93   \n",
       "\n",
       "   Outlet_Establishment_Year Outlet_Identifier Outlet_Location_Type  \\\n",
       "0                       1999            OUT049               Tier 1   \n",
       "1                       2009            OUT018               Tier 3   \n",
       "2                       1999            OUT049               Tier 1   \n",
       "3                       1998            OUT010               Tier 3   \n",
       "4                       1987            OUT013               Tier 3   \n",
       "\n",
       "  Outlet_Size        Outlet_Type source Item_Type_Combined  Outlet_Years  \n",
       "0      Medium  Supermarket Type1  train               Food            20  \n",
       "1      Medium  Supermarket Type2  train             Drinks            10  \n",
       "2      Medium  Supermarket Type1  train               Food            20  \n",
       "3       Small      Grocery Store  train               Food            21  \n",
       "4        High  Supermarket Type1  train     Non-Consumable            32  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since scikit-learn accepts only numerical variables, I converted all categories of nominal variables into numeric types."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\"\"\"Vectorising the data based on categories\n",
    "Labelencoder will change the different category in column values to respective unique\n",
    "number ex 0,1,2\n",
    "\n",
    "Onehotencoder will convert the 0,1,2 category to respective 3 columns with their\n",
    "binary values or dummy variables\n",
    "\n",
    "Dummy variable can also be made using pd.get_dummies()\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Item_Fat_Content', 'Item_Identifier', 'Item_MRP', 'Item_Outlet_Sales',\n",
       "       'Item_Type', 'Item_Visibility', 'Item_Weight',\n",
       "       'Outlet_Establishment_Year', 'Outlet_Identifier',\n",
       "       'Outlet_Location_Type', 'Outlet_Size', 'Outlet_Type', 'source',\n",
       "       'Item_Type_Combined', 'Outlet_Years'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.columns\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Item_Fat_Content',\n",
       " 'Outlet_Identifier',\n",
       " 'Outlet_Location_Type',\n",
       " 'Outlet_Size',\n",
       " 'Outlet_Type',\n",
       " 'Item_Type_Combined']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[i for i in categorical_columns if i not in['Item_Identifier','source','Item_Type',]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "#Feature scaling the category variable\n",
    "#label encoder only works when data dont have any missing values\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "#not using Item type i.e original col instaed using the converted one.\n",
    "\n",
    "var=[i for i in categorical_columns if i not in['Item_Identifier','source','Item_Type']]\n",
    "number=LabelEncoder() \n",
    "# first checking no columns have null or missing values.\n",
    "for i in var:\n",
    "    (data[i].isnull().unique())\n",
    "\n",
    "#converting to unique values \n",
    "for i in var:\n",
    "    data[i]=number.fit_transform(data[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Item_Identifier</th>\n",
       "      <th>Item_MRP</th>\n",
       "      <th>Item_Outlet_Sales</th>\n",
       "      <th>Item_Type</th>\n",
       "      <th>Item_Visibility</th>\n",
       "      <th>Item_Weight</th>\n",
       "      <th>Outlet_Establishment_Year</th>\n",
       "      <th>source</th>\n",
       "      <th>Outlet_Years</th>\n",
       "      <th>Item_Fat_Content_1</th>\n",
       "      <th>...</th>\n",
       "      <th>Outlet_Identifier_9</th>\n",
       "      <th>Outlet_Location_Type_1</th>\n",
       "      <th>Outlet_Location_Type_2</th>\n",
       "      <th>Outlet_Size_1</th>\n",
       "      <th>Outlet_Size_2</th>\n",
       "      <th>Outlet_Type_1</th>\n",
       "      <th>Outlet_Type_2</th>\n",
       "      <th>Outlet_Type_3</th>\n",
       "      <th>Item_Type_Combined_1</th>\n",
       "      <th>Item_Type_Combined_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>FDA15</td>\n",
       "      <td>249.8092</td>\n",
       "      <td>3735.1380</td>\n",
       "      <td>Dairy</td>\n",
       "      <td>0.016047</td>\n",
       "      <td>9.30</td>\n",
       "      <td>1999</td>\n",
       "      <td>train</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DRC01</td>\n",
       "      <td>48.2692</td>\n",
       "      <td>443.4228</td>\n",
       "      <td>Soft Drinks</td>\n",
       "      <td>0.019278</td>\n",
       "      <td>5.92</td>\n",
       "      <td>2009</td>\n",
       "      <td>train</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>FDN15</td>\n",
       "      <td>141.6180</td>\n",
       "      <td>2097.2700</td>\n",
       "      <td>Meat</td>\n",
       "      <td>0.016760</td>\n",
       "      <td>17.50</td>\n",
       "      <td>1999</td>\n",
       "      <td>train</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>FDX07</td>\n",
       "      <td>182.0950</td>\n",
       "      <td>732.3800</td>\n",
       "      <td>Fruits and Vegetables</td>\n",
       "      <td>0.017834</td>\n",
       "      <td>19.20</td>\n",
       "      <td>1998</td>\n",
       "      <td>train</td>\n",
       "      <td>21</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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",
       "      <th>4</th>\n",
       "      <td>NCD19</td>\n",
       "      <td>53.8614</td>\n",
       "      <td>994.7052</td>\n",
       "      <td>Household</td>\n",
       "      <td>0.009780</td>\n",
       "      <td>8.93</td>\n",
       "      <td>1987</td>\n",
       "      <td>train</td>\n",
       "      <td>32</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  Item_Identifier  Item_MRP  Item_Outlet_Sales              Item_Type  \\\n",
       "0           FDA15  249.8092          3735.1380                  Dairy   \n",
       "1           DRC01   48.2692           443.4228            Soft Drinks   \n",
       "2           FDN15  141.6180          2097.2700                   Meat   \n",
       "3           FDX07  182.0950           732.3800  Fruits and Vegetables   \n",
       "4           NCD19   53.8614           994.7052              Household   \n",
       "\n",
       "   Item_Visibility  Item_Weight  Outlet_Establishment_Year source  \\\n",
       "0         0.016047         9.30                       1999  train   \n",
       "1         0.019278         5.92                       2009  train   \n",
       "2         0.016760        17.50                       1999  train   \n",
       "3         0.017834        19.20                       1998  train   \n",
       "4         0.009780         8.93                       1987  train   \n",
       "\n",
       "   Outlet_Years  Item_Fat_Content_1          ...           \\\n",
       "0            20                   0          ...            \n",
       "1            10                   1          ...            \n",
       "2            20                   0          ...            \n",
       "3            21                   1          ...            \n",
       "4            32                   0          ...            \n",
       "\n",
       "   Outlet_Identifier_9  Outlet_Location_Type_1  Outlet_Location_Type_2  \\\n",
       "0                    1                       0                       0   \n",
       "1                    0                       0                       1   \n",
       "2                    1                       0                       0   \n",
       "3                    0                       0                       1   \n",
       "4                    0                       0                       1   \n",
       "\n",
       "   Outlet_Size_1  Outlet_Size_2  Outlet_Type_1  Outlet_Type_2  Outlet_Type_3  \\\n",
       "0              1              0              1              0              0   \n",
       "1              1              0              0              1              0   \n",
       "2              1              0              1              0              0   \n",
       "3              0              1              0              0              0   \n",
       "4              0              0              1              0              0   \n",
       "\n",
       "   Item_Type_Combined_1  Item_Type_Combined_2  \n",
       "0                     1                     0  \n",
       "1                     0                     0  \n",
       "2                     1                     0  \n",
       "3                     1                     0  \n",
       "4                     0                     1  \n",
       "\n",
       "[5 rows x 28 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#creating dummy variables\n",
    "\n",
    "# from sklearn.preprocessing import OneHotEncoder\n",
    "# onehotencoder= OneHotEncoder(categorical_features=[0])\n",
    "# for i in var:\n",
    "#     data[i]= onehotencoder.fit_transform(data[i]).toarray()\n",
    "\n",
    "data = pd.get_dummies(data,columns=var,drop_first=True)\n",
    "data.head()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Final step is to convert data back into train and test data sets. Its generally a good idea to export both of these as modified data sets so that they can be re-used for multiple sessions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Drop the columns which have been converted to different types:\n",
    "data.drop(['Item_Type','Outlet_Establishment_Year'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.7/site-packages/pandas/core/frame.py:3697: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  errors=errors)\n"
     ]
    }
   ],
   "source": [
    "#Divide into test and train:\n",
    "train = data.loc[data['source']==\"train\"]\n",
    "test = data.loc[data['source']==\"test\"]\n",
    "\n",
    "#Drop unnecessary columns:\n",
    "test.drop(['Item_Outlet_Sales','source'],axis=1,inplace=True)\n",
    "train.drop(['source'],axis=1,inplace=True)\n",
    "\n",
    "#Export files as modified versions:\n",
    "# train.to_csv(\"train_modified.csv\",index=False)\n",
    "# test.to_csv(\"test_modified.csv\",index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Building"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Baseline Model: the one which requires no predictive model and its like an informed guess. For instance, in this case lets predict the sales as the overall average sales."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.7/site-packages/pandas/core/indexing.py:1472: FutureWarning: \n",
      "Passing list-likes to .loc or [] with any missing label will raise\n",
      "KeyError in the future, you can use .reindex() as an alternative.\n",
      "\n",
      "See the documentation here:\n",
      "https://pandas.pydata.org/pandas-docs/stable/indexing.html#deprecate-loc-reindex-listlike\n",
      "  return self._getitem_tuple(key)\n"
     ]
    },
    {
     "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>Item_Identifier</th>\n",
       "      <th>Outlet_Identifier</th>\n",
       "      <th>Item_Outlet_Sales</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8523</th>\n",
       "      <td>FDW58</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2181.288914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8524</th>\n",
       "      <td>FDW14</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2181.288914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8525</th>\n",
       "      <td>NCN55</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2181.288914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8526</th>\n",
       "      <td>FDQ58</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2181.288914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8527</th>\n",
       "      <td>FDY38</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2181.288914</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Item_Identifier  Outlet_Identifier  Item_Outlet_Sales\n",
       "8523           FDW58                NaN        2181.288914\n",
       "8524           FDW14                NaN        2181.288914\n",
       "8525           NCN55                NaN        2181.288914\n",
       "8526           FDQ58                NaN        2181.288914\n",
       "8527           FDY38                NaN        2181.288914"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Mean based:\n",
    "mean_sales = train['Item_Outlet_Sales'].mean()\n",
    "\n",
    "#Define a dataframe with IDs for submission:\n",
    "base1 = test.loc[:,['Item_Identifier','Outlet_Identifier']]\n",
    "base1['Item_Outlet_Sales'] = mean_sales\n",
    "base1.head()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " Making baseline models helps in setting a benchmark. If your predictive algorithm is below this, there is something going seriously wrong and you should check your data.\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8523, 25)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "train.shape\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear Regression Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "y= train.loc[:,'Item_Outlet_Sales'].values\n",
    "x= train.loc[:, [x for x in train.columns if x not in ['Item_Outlet_Sales','Item_Identifier'] ]].values\n",
    "# we are not taking id columns as there is no use of identifiers\n",
    "\n",
    "#splitting the data of training and test data\n",
    "from sklearn.model_selection import train_test_split\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 1/3, random_state = 0)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Normalization\n",
    "\n",
    "Normalization is a technique often applied as part of data preparation for machine learning. The goal of normalization is to change the values of numeric columns in the dataset to a common scale, without distorting differences in the ranges of values. For machine learning, every dataset does not require normalization. It is required only when features have different ranges.\n",
    "\n",
    "\n",
    "For example, consider a data set containing two features, age(x1), and income(x2). Where age ranges from 0–100, while income ranges from 0–20,000 and higher. Income is about 1,000 times larger than age and ranges from 20,000–500,000. So, these two features are in very different ranges. When we do further analysis, like multivariate linear regression, for example, the attributed income will intrinsically influence the result more due to its larger value. But this doesn’t necessarily mean it is more important as a predictor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "ss= StandardScaler()\n",
    "x_train=ss.fit_transform(x_train)\n",
    "x_test=ss.transform(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LinearRegression, Ridge, Lasso"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "#making the regression object\n",
    "reg= LinearRegression(normalize=False)\n",
    "#here normalize True means,\n",
    "#the regressors X will be normalized before regression by subtracting the mean and dividing by the l2-norm\n",
    "# If you wish to standardize,\n",
    "# please use sklearn.preprocessing.StandardScaler before calling fit on an estimator with normalize=False."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None,\n",
       "         normalize=False)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reg.fit(x_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 9.59153623e+02, -5.23047301e+00, -9.68323215e+00,  2.02970671e+15,\n",
       "        1.92893211e+01, -2.09315222e+14, -3.95636390e+15, -4.08260835e+15,\n",
       "        1.39284615e+14, -4.46575044e+15, -4.13717944e+15, -4.30829054e+15,\n",
       "        2.00064045e+15, -7.53842244e+15,  1.00142055e+16,  1.82160187e+15,\n",
       "        1.41019557e+16, -4.01999077e+14, -1.33574061e+15, -4.73982932e+15,\n",
       "       -6.11027644e+15,  1.26230851e+01,  2.18005542e-01])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Predict training set:\n",
    "y_pred = reg.predict(x_test)\n",
    "# plt.plot(y_test,y_pred)\n",
    "# plt.show()\n",
    "# %matplotlib inline\n",
    "reg.coef_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cross Validation is a technique which involves reserving a particular sample of a dataset on which you do not train the model.\n",
    "\n",
    "the steps involved in cross validation:\n",
    "\n",
    "You reserve a sample data set\n",
    "Train the model using the remaining part of the dataset\n",
    "Use the reserve sample of the test (validation) set. This will help you in gauging the effectiveness of your model’s performance. If your model delivers a positive result on validation data, go ahead with the current model. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Perform cross-validation using The validation set approach:\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.metrics import mean_absolute_error, mean_squared_error\n",
    "cv_score = cross_val_score(reg, x_test, y_test, cv=20, scoring='neg_mean_squared_error')\n",
    "cv_score = np.sqrt(np.abs(cv_score))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "scoring metrcs have ['accuracy', 'adjusted_rand_score', 'average_precision', 'f1', 'f1_macro', 'f1_micro', 'f1_samples',\n",
    "                     'f1_weighted', 'neg_log_loss', 'neg_mean_absolute_error', 'neg_mean_squared_error', 'neg_median_absolute_error', 'precision', 'precision_macro', 'precision_micro', 'precision_samples', 'precision_weighted',\n",
    "                     'r2', 'recall', 'recall_macro', 'recall_micro', 'recall_samples', 'recall_weighted', 'roc_auc']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1195.89525291, 1069.92767392, 1168.78913239, 1184.25023693,\n",
       "       1106.01020209, 1138.0360716 , 1061.66830811, 1268.82238229,\n",
       "       1123.60499819, 1107.84272469, 1154.31433172, 1009.69450792,\n",
       "       1098.51987177, 1254.57880108, 1281.08165838, 1288.11007648,\n",
       "       1161.26494306, 1097.77948509,  990.36992371, 1048.23674698])"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 9.59153623e+02, -5.23047301e+00, -9.68323215e+00,  2.02970671e+15,\n",
       "        1.92893211e+01, -2.09315222e+14, -3.95636390e+15, -4.08260835e+15,\n",
       "        1.39284615e+14, -4.46575044e+15, -4.13717944e+15, -4.30829054e+15,\n",
       "        2.00064045e+15, -7.53842244e+15,  1.00142055e+16,  1.82160187e+15,\n",
       "        1.41019557e+16, -4.01999077e+14, -1.33574061e+15, -4.73982932e+15,\n",
       "       -6.11027644e+15,  1.26230851e+01,  2.18005542e-01])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reg.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "23"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(reg.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Model Report\n",
      "Root mean squared error : 1140\n",
      "CV Score : Mean - 1140 | Std - 84.54 | Min - 990.4 | Max - 1288\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x12322dc88>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "    #Print model report:\n",
    "    print (\"\\nModel Report\")\n",
    "    #Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors:\n",
    "    print (\"Root mean squared error : %.4g\" % np.sqrt(mean_squared_error(y_test, y_pred)))\n",
    "    print (\"CV Score : Mean - %.4g | Std - %.4g | Min - %.4g | Max - %.4g\" % (np.mean(cv_score),np.std(cv_score),np.min(cv_score),np.max(cv_score)))\n",
    "    coef1 = pd.Series(reg.coef_, [x for x in train.columns if x not in ['Item_Outlet_Sales','Item_Identifier', 'Outlet_Identifier']]).sort_values()\n",
    "    coef1.plot(kind='bar', title='Model Coefficients')\n",
    "    # The absolute value of the correlation coefficient denotes the strength of the relationship.\n",
    "    #Since  absolute correlation is very high it means that the relationship is strong"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluation Metrics\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluation metrics for classification problems, such as accuracy, are not useful for regression problems.\n",
    "Instead, we need evaluation metrics designed for comparing continuous values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE: 849.3315\n",
      "MSE: 1298986.3663\n",
      "Root mean squared error : 1139.7308\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Mean Absolute Error (MAE) is the mean of the absolute value of the errors:\n",
    "print(\"MAE: %.4f\" % (mean_absolute_error(y_test, y_pred)))\n",
    "\n",
    "# Mean Squared Error (MSE) is the mean of the squared errors:\n",
    "print(\"MSE: %.4f\" % (mean_squared_error(y_test, y_pred)))\n",
    "\n",
    "#Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors, less rmse means good model:\n",
    "print (\"Root mean squared error : %.4f\" % np.sqrt(mean_squared_error(y_test, y_pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Comparing these metrics:\n",
    "\n",
    "MAE is the easiest to understand, because it's the average error.\n",
    "MSE is more popular than MAE, because MSE \"punishes\" larger errors.\n",
    "RMSE is even more popular than MSE, because RMSE is interpretable in the \"y\" units and Easier to put in context as it's the same units as our response variable"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Residuals refer to the error values of the model. Therefore lower residuals are desired."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ‘Ridge Regression‘ and ‘Lasso Regression‘ the most fundamental regularization techniques."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ridge Regression is a technique used when the data suffers from multicollinearity ( independent variables are highly correlated).In multicollinearity, even though the least squares estimates (OLS) are unbiased, their variances are large which deviates the observed value far from the true value. By adding a degree of bias to the regression estimates, ridge regression reduces the standard errors.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "y=a+b*x+e (error term),  [error term is the value needed to correct for a prediction error between the observed and predicted value]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "prediction errors can be decomposed into two sub components. First is due to the biased and second is due to the variance. Prediction error can occur due to any one of these two or both components. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ridge regression solves the multicollinearity problem through shrinkage parameter λ (lambda), which adds the penalities to independent parameter to shrink the parameter to have a very low variance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is Regularization which basically adds the penalty as model complexity increases. Regularization parameter (lambda) penalizes all the parameters except intercept so that model generalizes the data and won’t overfit.\n",
    "\n",
    "Ridge follows L2 regularization.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "URL: https://towardsdatascience.com/l1-and-l2-regularization-methods-ce25e7fc831c\n",
    "In order to create less complex (parsimonious) model when you have a large number of features in your dataset, some of the Regularization techniques used to address over-fitting and feature selection are:\n",
    "\n",
    "1. L1 Regularization : Lasso Regression (Least Absolute Shrinkage and Selection Operator) adds “absolute value of magnitude” of coefficient as penalty term to the loss function. if lambda is zero then we will get back OLS whereas very large value will make coefficients zero hence it will under-fit.\n",
    "\n",
    "2. L2 Regularization or Ridge : Ridge regression adds “squared magnitude” of coefficient as penalty term to the loss function. if lambda is zero then you can imagine we get back OLS. However, if lambda is very large then it will add too much weight and it will lead to under-fitting. Having said that it’s important how lambda is chosen. This technique works very well to avoid over-fitting issue.\n",
    "\n",
    "A regression model that uses L1 regularization technique is called Lasso Regression and model which uses L2 is called Ridge Regression.\n",
    "\n",
    "The key difference between these techniques is that Lasso shrinks the less important feature’s coefficient to zero thus, removing some feature altogether. So, this works well for feature selection in case we have a huge number of features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import Ridge , Lasso"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "r= Ridge(alpha=0.2, normalize=False)\n",
    "## higher the alpha value, more restriction on the coefficients; low alpha > more generalization, coefficients are barely restricted"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "r.fit(x_train,y_train)\n",
    "yr= r.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_score = cross_val_score(r, x_test, y_test, cv=20, scoring='neg_mean_squared_error')\n",
    "cv_score = np.sqrt(np.abs(cv_score))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1196.00814656, 1070.18054945, 1169.04380886, 1183.86922046,\n",
       "       1106.1196852 , 1138.82351774, 1059.70794223, 1268.48552417,\n",
       "       1122.60251239, 1108.25636054, 1153.21023716, 1010.24232545,\n",
       "       1097.90490391, 1254.05306905, 1280.89897897, 1286.96820744,\n",
       "       1161.07807594, 1097.78912009,  991.75380028, 1048.27005118])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 959.73691474,   -4.56428194,   -6.91915915,   57.51398497,\n",
       "         16.77646125,  133.81548317,  123.9278797 ,   93.08732061,\n",
       "       -136.6462329 ,  329.07369931,  120.59235913,   66.01215763,\n",
       "        204.14717583,  -89.75959214,  205.52597474, -203.73816031,\n",
       "        220.13678505, -292.3599174 ,  365.74257899,   93.08732061,\n",
       "        329.07369932,   16.71389488,    4.52469458])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "23"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(r.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Model Report\n",
      "Root mean squared error : 1140\n",
      "CV Score : Mean - 1140 | Std - 84.28 | Min - 991.8 | Max - 1287\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1273efc50>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "    #Print model report:\n",
    "    print (\"\\nModel Report\")\n",
    "    #Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors:\n",
    "    print (\"Root mean squared error : %.4g\" % np.sqrt(mean_squared_error(y_test, yr)))\n",
    "    print (\"CV Score : Mean - %.4g | Std - %.4g | Min - %.4g | Max - %.4g\" % (np.mean(cv_score),np.std(cv_score),np.min(cv_score),np.max(cv_score)))\n",
    "    coef1 = pd.Series(r.coef_, [x for x in train.columns if x not in ['Item_Outlet_Sales','Item_Identifier', 'Outlet_Identifier']]).sort_values()\n",
    "    coef1.plot(kind='bar', title='Model Coefficients')\n",
    "    # The absolute value of the correlation coefficient denotes the strength of the relationship.\n",
    "    #Since  absolute correlation is very high it means that the relationship is strong"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here , magnitude of coefficients i slighty better tthan linear regression and also score is similar"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE: 849.4949\n",
      "MSE: 1299125.4612\n",
      "Root mean squared error : 1139.7918\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Mean Absolute Error (MAE) is the mean of the absolute value of the errors:\n",
    "print(\"MAE: %.4f\" % (mean_absolute_error(y_test, yr)))\n",
    "\n",
    "# Mean Squared Error (MSE) is the mean of the squared errors:\n",
    "print(\"MSE: %.4f\" % (mean_squared_error(y_test, yr)))\n",
    "\n",
    "#Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors, less rmse means good model:\n",
    "print (\"Root mean squared error : %.4f\" % np.sqrt(mean_squared_error(y_test, yr)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "l=Lasso(alpha=0.5, normalize=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/coordinate_descent.py:492: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations. Fitting data with very small alpha may cause precision problems.\n",
      "  ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Lasso(alpha=0.5, copy_X=True, fit_intercept=True, max_iter=1000,\n",
       "   normalize=False, positive=False, precompute=False, random_state=None,\n",
       "   selection='cyclic', tol=0.0001, warm_start=False)"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l.fit(x_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "yl= l.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1139.8854077311082"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sqrt(mean_squared_error(y_test,yl))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_score = cross_val_score(r, x_test, y_test, cv=20, scoring='neg_mean_squared_error')\n",
    "cv_score = np.sqrt(np.abs(cv_score))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Model Report\n",
      "Root mean squared error : 1140\n",
      "CV Score : Mean - 1140 | Std - 84.28 | Min - 991.8 | Max - 1287\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x12bece390>"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "    #Print model report:\n",
    "    print (\"\\nModel Report\")\n",
    "    #Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors:\n",
    "    print (\"Root mean squared error : %.4g\" % np.sqrt(mean_squared_error(y_test, yl)))\n",
    "    print (\"CV Score : Mean - %.4g | Std - %.4g | Min - %.4g | Max - %.4g\" % (np.mean(cv_score),np.std(cv_score),np.min(cv_score),np.max(cv_score)))\n",
    "    coef1 = pd.Series(l.coef_, [x for x in train.columns if x not in ['Item_Outlet_Sales','Item_Identifier', 'Outlet_Identifier']]).sort_values()\n",
    "    coef1.plot(kind='bar', title='Model Coefficients')\n",
    "    # The absolute value of the correlation coefficient denotes the strength of the relationship.\n",
    "    #Since  absolute correlation is very high it means that the relationship is strong"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "#applyig decision Tree regression.\n",
    "\n",
    "# Applying the model\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "decreg= DecisionTreeRegressor(max_depth=8, min_samples_leaf=50,random_state= 0)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeRegressor(criterion='mse', max_depth=8, max_features=None,\n",
       "           max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
       "           min_impurity_split=None, min_samples_leaf=50,\n",
       "           min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "           presort=False, random_state=0, splitter='best')"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decreg.fit(x_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<bound method RegressorMixin.score of DecisionTreeRegressor(criterion='mse', max_depth=8, max_features=None,\n",
       "           max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
       "           min_impurity_split=None, min_samples_leaf=50,\n",
       "           min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "           presort=False, random_state=0, splitter='best')>"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decreg.score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "ydec= decreg.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5.49534914e-01, 9.88473862e-03, 3.05233043e-03, 1.56665846e-03,\n",
       "       0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 7.78041484e-02,\n",
       "       0.00000000e+00, 4.27471851e-02, 0.00000000e+00, 0.00000000e+00,\n",
       "       1.33705873e-03, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "       1.13512768e-03, 1.33918063e-03, 1.67683928e-01, 2.23790372e-02,\n",
       "       1.21498631e-01, 3.70618666e-05, 0.00000000e+00])"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decreg.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "cv_score = cross_val_score(decreg, x_test, y_test, cv=20, scoring='neg_mean_squared_error')\n",
    "cv_score = np.sqrt(np.abs(cv_score))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Model Report\n",
      "Root mean squared error : 1110\n",
      "CV Score : Mean - 1118 | Std - 95.13 | Min - 981.9 | Max - 1347\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x12bfeb828>"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "    #Print model report:\n",
    "    print (\"\\nModel Report\")\n",
    "    #Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors:\n",
    "    print (\"Root mean squared error : %.4g\" % np.sqrt(mean_squared_error(y_test, ydec)))\n",
    "    print (\"CV Score : Mean - %.4g | Std - %.4g | Min - %.4g | Max - %.4g\" % (np.mean(cv_score),np.std(cv_score),np.min(cv_score),np.max(cv_score)))\n",
    "    coef1 = pd.Series(decreg.feature_importances_, [x for x in train.columns if x not in ['Item_Outlet_Sales','Item_Identifier', 'Outlet_Identifier']]).sort_values()\n",
    "    coef1.plot(kind='bar', title='Model Coefficients')\n",
    "    # The absolute value of the correlation coefficient denotes the strength of the relationship.\n",
    "    #Since  absolute correlation is very high it means that the relationship is strong"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now considering the top 5 features for the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Model Report\n",
      "Root mean squared error : 1107\n",
      "CV Score : Mean - 1109 | Std - 85.03 | Min - 962.1 | Max - 1317\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x12c1607b8>"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "top5= ['Item_MRP','Outlet_Type_1','Outlet_Years','Outlet_Type_3','Outlet_Identifier_5']\n",
    "y= train.loc[:,'Item_Outlet_Sales'].values\n",
    "x= train.loc[:, top5].values\n",
    "# we are not taking id columns as there is no use of identifiers\n",
    "\n",
    "#splitting the data of training and test data\n",
    "from sklearn.model_selection import train_test_split\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 1/3, random_state = 0)\n",
    "\n",
    "\n",
    "decreg= DecisionTreeRegressor(max_depth=8, min_samples_leaf=50,random_state= 0)\n",
    "decreg.fit(x_train, y_train)\n",
    "\n",
    "\n",
    "ydec= decreg.predict(x_test)\n",
    "cv_score = cross_val_score(decreg, x_test, y_test, cv=20, scoring='neg_mean_squared_error')\n",
    "cv_score = np.sqrt(np.abs(cv_score))\n",
    "print (\"\\nModel Report\")\n",
    "#Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors:\n",
    "print (\"Root mean squared error : %.4g\" % np.sqrt(mean_squared_error(y_test, ydec)))\n",
    "print (\"CV Score : Mean - %.4g | Std - %.4g | Min - %.4g | Max - %.4g\" % (np.mean(cv_score),np.std(cv_score),np.min(cv_score),np.max(cv_score)))\n",
    "coef1 = pd.Series(decreg.feature_importances_, top5).sort_values()\n",
    "coef1.plot(kind='bar', title='Model Coefficients')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random Forest"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Random Forest is a versatile machine learning method capable of performing both regression and classification tasks. It also undertakes dimensional reduction methods, treats missing values, outlier values and other essential steps of data exploration, and does a fairly good job. It is a type of ensemble learning method, where a group of weak models combine to form a powerful model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Model Report\n",
      "Root mean squared error : 1095\n",
      "CV Score : Mean - 1109 | Std - 88.32 | Min - 949.3 | Max - 1294\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x122b94ac8>"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "regrf= RandomForestRegressor(n_estimators=200,max_depth=8,min_samples_leaf=50,n_jobs=4)\n",
    "\n",
    "regrf.fit(x_train,y_train)\n",
    "\n",
    "regrf.feature_importances_\n",
    "\n",
    "yrf= regrf.predict(x_test)\n",
    "cv_score = cross_val_score(regrf, x_test, y_test, cv=20, scoring='neg_mean_squared_error')\n",
    "cv_score = np.sqrt(np.abs(cv_score))\n",
    "print (\"\\nModel Report\")\n",
    "#Root Mean Squared Error (RMSE) is the square root of the mean of the squared errors:\n",
    "print (\"Root mean squared error : %.4g\" % np.sqrt(mean_squared_error(y_test, yrf)))\n",
    "print (\"CV Score : Mean - %.4g | Std - %.4g | Min - %.4g | Max - %.4g\" % (np.mean(cv_score),np.std(cv_score),np.min(cv_score),np.max(cv_score)))\n",
    "coef1 = pd.Series(decreg.feature_importances_, top5).sort_values()\n",
    "coef1.plot(kind='bar', title='Model Coefficients')"
   ]
  },
  {
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}