Load libraries and data:

The data will be the one exported from the 'exploration_and_feature_engineering' file



In [1]:

    
import pandas as pd
import numpy as np
%matplotlib inline
from matplotlib.pylab import rcParams
rcParams['figure.figsize'] = 12, 8



In [2]:

    
train = pd.read_csv("train_modified.csv")
test = pd.read_csv("test_modified.csv")



In [3]:

    
print train.shape
train.dtypes









    



(8523, 34)






    Out[3]:





Item_Identifier               object
Item_MRP                     float64
Item_Outlet_Sales            float64
Item_Visibility              float64
Item_Weight                  float64
Outlet_Identifier             object
Item_Visibility_MeanRatio    float64
Outlet_Years                   int64
Item_Fat_Content_0           float64
Item_Fat_Content_1           float64
Item_Fat_Content_2           float64
Outlet_Location_Type_0       float64
Outlet_Location_Type_1       float64
Outlet_Location_Type_2       float64
Outlet_Size_0                float64
Outlet_Size_1                float64
Outlet_Size_2                float64
Outlet_Type_0                float64
Outlet_Type_1                float64
Outlet_Type_2                float64
Outlet_Type_3                float64
Item_Type_Combined_0         float64
Item_Type_Combined_1         float64
Item_Type_Combined_2         float64
Outlet_0                     float64
Outlet_1                     float64
Outlet_2                     float64
Outlet_3                     float64
Outlet_4                     float64
Outlet_5                     float64
Outlet_6                     float64
Outlet_7                     float64
Outlet_8                     float64
Outlet_9                     float64
dtype: object

Baseline models:



In [4]:

    
#Mean based:
mean_sales = train['Item_Outlet_Sales'].mean()

#Define a dataframe with IDs for submission:
base1 = test[['Item_Identifier','Outlet_Identifier']]
base1['Item_Outlet_Sales'] = mean_sales

#Export submission file
base1.to_csv("alg0.csv",index=False)









    



/Users/aarshay/anaconda/lib/python2.7/site-packages/ipykernel/__main__.py:6: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy

Function to fit and generate submission file:



In [5]:

    
#Define target and ID columns:
target = 'Item_Outlet_Sales'
IDcol = ['Item_Identifier','Outlet_Identifier']
from sklearn import cross_validation, metrics
def modelfit(alg, dtrain, dtest, predictors, target, IDcol, filename):
    #Fit the algorithm on the data
    alg.fit(dtrain[predictors], dtrain[target])
        
    #Predict training set:
    dtrain_predictions = alg.predict(dtrain[predictors])

    #Perform cross-validation:
    cv_score = cross_validation.cross_val_score(alg, dtrain[predictors], dtrain[target], cv=20, scoring='mean_squared_error')
    cv_score = np.sqrt(np.abs(cv_score))
    
    #Print model report:
    print "\nModel Report"
    print "RMSE : %.4g" % np.sqrt(metrics.mean_squared_error(dtrain[target].values, dtrain_predictions))
    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))
    
    #Predict on testing data:
    dtest[target] = alg.predict(dtest[predictors])
    
    #Export submission file:
    IDcol.append(target)
    submission = pd.DataFrame({ x: dtest[x] for x in IDcol})
    submission.to_csv(filename, index=False)

Linear Regression Model:



In [6]:

    
from sklearn.linear_model import LinearRegression, Ridge



In [7]:

    
predictors = [x for x in train.columns if x not in [target]+IDcol]
# print predictors
alg1 = LinearRegression(normalize=True)
modelfit(alg1, train, test, predictors, target, IDcol, 'alg1.csv')
coef1 = pd.Series(alg1.coef_, predictors).sort_values()
coef1.plot(kind='bar', title='Model Coefficients')









    



Model Report
RMSE : 1128
CV Score : Mean - 1129 | Std - 44.16 | Min - 1074 | Max - 1218






    Out[7]:





<matplotlib.axes._subplots.AxesSubplot at 0x10b4f3650>

Ridge Regression Model:



In [8]:

    
predictors = [x for x in train.columns if x not in [target]+IDcol]
alg2 = Ridge(alpha=0.05,normalize=True)
modelfit(alg2, train, test, predictors, target, IDcol, 'alg2.csv')
coef2 = pd.Series(alg2.coef_, predictors).sort_values()
coef2.plot(kind='bar', title='Model Coefficients')









    



Model Report
RMSE : 1129
CV Score : Mean - 1130 | Std - 44.6 | Min - 1076 | Max - 1217






    Out[8]:





<matplotlib.axes._subplots.AxesSubplot at 0x10cf61450>

Decision Tree Model:



In [9]:

    
from sklearn.tree import DecisionTreeRegressor



In [10]:

    
predictors = [x for x in train.columns if x not in [target]+IDcol]
alg3 = DecisionTreeRegressor(max_depth=15, min_samples_leaf=100)
modelfit(alg3, train, test, predictors, target, IDcol, 'alg3.csv')
coef3 = pd.Series(alg3.feature_importances_, predictors).sort_values(ascending=False)
coef3.plot(kind='bar', title='Feature Importances')









    



Model Report
RMSE : 1058
CV Score : Mean - 1091 | Std - 45.42 | Min - 1003 | Max - 1186






    Out[10]:





<matplotlib.axes._subplots.AxesSubplot at 0x10d687250>



In [11]:

    
predictors = ['Item_MRP','Outlet_Type_0','Outlet_5','Outlet_Years']
alg4 = DecisionTreeRegressor(max_depth=8, min_samples_leaf=150)
modelfit(alg4, train, test, predictors, target, IDcol, 'alg4.csv')
coef4 = pd.Series(alg4.feature_importances_, predictors).sort_values(ascending=False)
coef4.plot(kind='bar', title='Feature Importances')









    



Model Report
RMSE : 1071
CV Score : Mean - 1096 | Std - 43.3 | Min - 1027 | Max - 1172






    Out[11]:





<matplotlib.axes._subplots.AxesSubplot at 0x10da94b10>

Random Forest Model:

Note: random forest models are not 100% replicable. So the outputs might differ very slightly but should be around the ballpark.



In [12]:

    
from sklearn.ensemble import RandomForestRegressor



In [13]:

    
predictors = [x for x in train.columns if x not in [target]+IDcol]
alg5 = RandomForestRegressor(n_estimators=200,max_depth=5, min_samples_leaf=100,n_jobs=4)
modelfit(alg5, train, test, predictors, target, IDcol, 'alg5.csv')
coef5 = pd.Series(alg5.feature_importances_, predictors).sort_values(ascending=False)
coef5.plot(kind='bar', title='Feature Importances')









    



Model Report
RMSE : 1073
CV Score : Mean - 1084 | Std - 43.79 | Min - 1020 | Max - 1160






    Out[13]:





<matplotlib.axes._subplots.AxesSubplot at 0x10df14d10>



In [14]:

    
predictors = [x for x in train.columns if x not in [target]+IDcol]
alg6 = RandomForestRegressor(n_estimators=400,max_depth=6, min_samples_leaf=100,n_jobs=4)
modelfit(alg6, train, test, predictors, target, IDcol, 'alg6.csv')
coef6 = pd.Series(alg6.feature_importances_, predictors).sort_values(ascending=False)
coef6.plot(kind='bar', title='Feature Importances')









    



Model Report
RMSE : 1068
CV Score : Mean - 1083 | Std - 43.77 | Min - 1020 | Max - 1162






    Out[14]:





<matplotlib.axes._subplots.AxesSubplot at 0x10df12450>