Python绘制KS曲线的实现方法

时间:2022-11-08 17:06:27

python实现KS曲线,相关使用方法请参考上篇博客-R语言实现KS曲线

代码如下:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
####################### PlotKS ##########################
def PlotKS(preds, labels, n, asc):
  
  # preds is score: asc=1
  # preds is prob: asc=0
  
  pred = preds # 预测值
  bad = labels # 取1为bad, 0为good
  ksds = DataFrame({'bad': bad, 'pred': pred})
  ksds['good'] = 1 - ksds.bad
  
  if asc == 1:
    ksds1 = ksds.sort_values(by=['pred', 'bad'], ascending=[True, True])
  elif asc == 0:
    ksds1 = ksds.sort_values(by=['pred', 'bad'], ascending=[False, True])
  ksds1.index = range(len(ksds1.pred))
  ksds1['cumsum_good1'] = 1.0*ksds1.good.cumsum()/sum(ksds1.good)
  ksds1['cumsum_bad1'] = 1.0*ksds1.bad.cumsum()/sum(ksds1.bad)
  
  if asc == 1:
    ksds2 = ksds.sort_values(by=['pred', 'bad'], ascending=[True, False])
  elif asc == 0:
    ksds2 = ksds.sort_values(by=['pred', 'bad'], ascending=[False, False])
  ksds2.index = range(len(ksds2.pred))
  ksds2['cumsum_good2'] = 1.0*ksds2.good.cumsum()/sum(ksds2.good)
  ksds2['cumsum_bad2'] = 1.0*ksds2.bad.cumsum()/sum(ksds2.bad)
  
  # ksds1 ksds2 -> average
  ksds = ksds1[['cumsum_good1', 'cumsum_bad1']]
  ksds['cumsum_good2'] = ksds2['cumsum_good2']
  ksds['cumsum_bad2'] = ksds2['cumsum_bad2']
  ksds['cumsum_good'] = (ksds['cumsum_good1'] + ksds['cumsum_good2'])/2
  ksds['cumsum_bad'] = (ksds['cumsum_bad1'] + ksds['cumsum_bad2'])/2
  
  # ks
  ksds['ks'] = ksds['cumsum_bad'] - ksds['cumsum_good']
  ksds['tile0'] = range(1, len(ksds.ks) + 1)
  ksds['tile'] = 1.0*ksds['tile0']/len(ksds['tile0'])
  
  qe = list(np.arange(0, 1, 1.0/n))
  qe.append(1)
  qe = qe[1:]
  
  ks_index = Series(ksds.index)
  ks_index = ks_index.quantile(q = qe)
  ks_index = np.ceil(ks_index).astype(int)
  ks_index = list(ks_index)
  
  ksds = ksds.loc[ks_index]
  ksds = ksds[['tile', 'cumsum_good', 'cumsum_bad', 'ks']]
  ksds0 = np.array([[0, 0, 0, 0]])
  ksds = np.concatenate([ksds0, ksds], axis=0)
  ksds = DataFrame(ksds, columns=['tile', 'cumsum_good', 'cumsum_bad', 'ks'])
  
  ks_value = ksds.ks.max()
  ks_pop = ksds.tile[ksds.ks.idxmax()]
  print ('ks_value is ' + str(np.round(ks_value, 4)) + ' at pop = ' + str(np.round(ks_pop, 4)))
  
  # chart
  plt.plot(ksds.tile, ksds.cumsum_good, label='cum_good',
             color='blue', linestyle='-', linewidth=2)
             
  plt.plot(ksds.tile, ksds.cumsum_bad, label='cum_bad',
            color='red', linestyle='-', linewidth=2)
            
  plt.plot(ksds.tile, ksds.ks, label='ks',
          color='green', linestyle='-', linewidth=2)
            
  plt.axvline(ks_pop, color='gray', linestyle='--')
  plt.axhline(ks_value, color='green', linestyle='--')
  plt.axhline(ksds.loc[ksds.ks.idxmax(), 'cumsum_good'], color='blue', linestyle='--')
  plt.axhline(ksds.loc[ksds.ks.idxmax(),'cumsum_bad'], color='red', linestyle='--')
  plt.title('KS=%s ' %np.round(ks_value, 4) +
        'at Pop=%s' %np.round(ks_pop, 4), fontsize=15)
  
 
  return ksds
####################### over ##########################

作图效果如下:

Python绘制KS曲线的实现方法

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持服务器之家。

原文链接:https://www.cnblogs.com/bigdatafengkong/p/9079093.html