In [1]:

    

import pandas as pd


    

In [3]:

    

sub1 = pd.read_csv('final.csv')
sub2 = pd.read_csv('xgb_final.csv')


    

In [7]:

    

sub1.shape, sub2.shape


    

    
Out[7]:





((2345796, 7), (2345796, 6))

In [6]:

    

sub1 = sub1.drop('test_id',axis=1)
sub2 = sub2.drop('test_id',axis=1)


    

In [18]:

    

dup = (sub1.fold1 + sub1.fold2 + sub1.fold3 + sub1.fold4 + sub1.fold5 + sub1.fold6 + sub1.fold7 + sub2.fold0 + sub2.fold1 + sub2.fold2 + sub2.fold3 +
      sub2.fold4 + sub2.fold5)/13


    

In [19]:

    

import matplotlib.pyplot as plt
plt.hist(dup,bins=50)
plt.show()


    

In [3]:

    

sub = pd.read_csv('sub_av.csv')


    

In [23]:

    

sub.is_duplicate = dup


    

In [24]:

    

sub.head()


    

    
Out[24]:







  

    

      

      
is_duplicate
      
test_id
    
  
  

    

      
0
      
0.000206
      
0
    
    

      
1
      
0.130279
      
1
    
    

      
2
      
0.188597
      
2
    
    

      
3
      
0.000090
      
3
    
    

      
4
      
0.075666
      
4
    
  

In [25]:

    

sub.to_csv('sub_final.csv',index = False)
! rm -rf test.zip
! zip -r test.zip sub_final.csv


    

    




  adding: sub_final.csv (deflated 59%)