

In [1]:

    
import numpy as np
import pandas as pd
import scipy as sp
import lightgbm as lgb
import gc



In [5]:

    
import matplotlib.pyplot as plt
%matplotlib inline



In [2]:

    
from sklearn.model_selection import train_test_split

Evaluate the feature importance



In [42]:

    
# load the lightgbm model
bst1 = lgb.Booster(model_file='../models/lightGBM_morefeatures_80_10.txt')



In [43]:

    
plt.rcParams["figure.figsize"] = 9, 7



In [44]:

    
lgb.plot_importance(bst1, max_num_features= 30)









    Out[44]:





<matplotlib.axes._subplots.AxesSubplot at 0x1938ecf8>

Select/Test different threshold to generate the predictions



In [45]:

    
prob_preds_df = pd.read_csv("../data/processed/lightGBM_morefeatures_prob_preds.csv")



In [46]:

    
prob_preds_df.head()



In [47]:

    
prob_preds_df.shape









    Out[47]:





(4833292, 3)



In [60]:

    
ax = prob_preds_df['pred'].plot.hist(bins = 30)
ax.set_xlabel("probability")









    Out[60]:





<matplotlib.text.Text at 0x31cfc5c0>



In [48]:

    
def generate_submission(df_test, test_orders_ids, file_name, threshold = 0.2, single_thres = True):
    """function to generate label predictions submission format"""
    if single_thres:
        TRESHOLD = threshold

        d = dict()
        for row in df_test.itertuples():
            if row.pred > TRESHOLD:
                try:
                    d[row.order_id] += ' ' + str(row.product_id)
                except:
                    d[row.order_id] = str(row.product_id)

        for order in test_orders_ids:
            if order not in d:
                d[order] = 'None'

        sub = pd.DataFrame.from_dict(d, orient='index')
        sub.reset_index(inplace=True)
        sub.columns = ['order_id', 'products']
        sub.to_csv(file_name, index=False)
    else:
        pass



In [49]:

    
orders_df = pd.read_csv("../data/raw/orders.csv")



In [50]:

    
test_orders_ids = orders_df[orders_df.eval_set == "test"].order_id



In [51]:

    
generate_submission(prob_preds_df, test_orders_ids, "../models/lightGBM_morefeatures_preds_30%thr.csv", threshold = 0.3, single_thres = True)



In [54]:

    
generate_submission(prob_preds_df, test_orders_ids, "../models/lightGBM_morefeatures_preds_25%thr.csv", threshold = 0.25, single_thres = True)



In [55]:

    
generate_submission(prob_preds_df, test_orders_ids, "../models/lightGBM_morefeatures_preds_15%thr.csv", threshold = 0.15, single_thres = True)



In [56]:

    
generate_submission(prob_preds_df, test_orders_ids, "../models/lightGBM_morefeatures_preds_10%thr.csv", threshold = 0.1, single_thres = True)



In [ ]:

	order_id	pred	product_id
0	2774568	0.165027	17668
1	2774568	0.163123	39190
2	2774568	0.123430	44683
3	2774568	0.381907	21903
4	2774568	0.152347	14992