# Deleting outliers as identified by EDA
train.drop(train[(train['grlivarea'] > 4000)
& (train['saleprice'] < 300000)].index,
inplace=True)
train_objs_num = len(train)
y = train['saleprice']
dataset = pd.concat(objs=[train.drop(columns=['saleprice']), test],
axis=0,
ignore_index=True)
all_data = dataset.copy()
all_data.shape
import Clean, Feature_Engineering, Simple_Stacking
all_data = Clean.model(all_data, train_objs_num)
all_data = Feature_Engineering.model(all_data)
## Apply Log tranformation to target variable as it is right skewed
y = np.log1p(y)
f_train = all_data[:train_objs_num]
f_test = all_data[train_objs_num:]
predictions = Simple_Stacking.model(f_train, y, f_test)
final_predictions = predictions
submission = pd.DataFrame({
'Id': sub_index,
'SalePrice': final_predictions.astype(float)
})
