标签：ima span max 概率分布 info and imp das com
代码实现：
data_test_2 = {‘gd‘:[1,1,1,1,1,1,0,0,0,0,0,0,0],‘score‘:[1,2,0,2,2,7,4,5,4,0,4,18,np.nan]}
data_test_2 = pd.DataFrame(data_test_2)
def ks_calc_cross(data,var_col,y_col):
‘‘‘
功能: 计算KS值，输出对应分割点和累计分布
输入值:-
输出值:
‘ks‘: KS值，‘crossdens‘: 好坏人累积概率分布以及其差值gap
‘‘‘
ks_df1 = pd.crosstab(data[var_col],data[y_col])
print(ks_df1)
print(ks_df1.cumsum(axis=0))
print(ks_df1.sum())
ks_df2 = ks_df1.cumsum(axis=0) / ks_df1.sum()
print(ks_df2)
ks_df2[‘gap‘] = abs(ks_df2[0] - ks_df2[1])
ks = ks_df2[‘gap‘].max()
return ks
ks_value = ks_calc_cross(data_test_2, ‘score‘, ‘gd‘)
print(ks_value)
gd 0 1 score 0.0 1 1 1.0 0 1 2.0 0 3 4.0 3 0 5.0 1 0 7.0 0 1 18.0 1 0 gd 0 1 score 0.0 1 1 1.0 1 2 2.0 1 5 4.0 4 5 5.0 5 5 7.0 5 6 18.0 6 6 gd 0 6 1 6 dtype: int64 gd 0 1 score 0.0 0.166667 0.166667 1.0 0.166667 0.333333 2.0 0.166667 0.833333 4.0 0.666667 0.833333 5.0 0.833333 0.833333 7.0 0.833333 1.000000 18.0 1.000000 1.000000 0.666666666667
庖丁解牛：
import pandas as pd import numpy as np data_test_2 = {‘gd‘:[1,1,1,1,1,1,0,0,0,0,0,0,0],‘score‘:[1,2,0,2,2,7,4,5,4,0,4,18,np.nan]} data_test_2 = pd.DataFrame(data_test_2) data_test_2
ks_df1 = pd.crosstab(data_test_2[‘score‘],data_test_2[‘gd‘]) ks_df1
ks_df1.cumsum(axis=0)
ks_df1.sum()
ks_df2 = ks_df1.cumsum(axis=0) / ks_df1.sum()
ks_df2
ks_df2[‘gap‘] = abs(ks_df2[0] - ks_df2[1]) ks_df2
ks_df2[‘gap‘].max()
标签：ima span max 概率分布 info and imp das com
原文地址：https://www.cnblogs.com/wqbin/p/11146836.html
