KS曲线python实现

实战

# 导入第三方模块
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt# 自定义绘制ks曲线的函数
def plot_ks(y_test, y_score, positive_flag):# 对y_test重新设置索引y_test.index = np.arange(len(y_test))# 构建目标数据集target_data = pd.DataFrame({'y_test':y_test, 'y_score':y_score})# 按y_score降序排列target_data.sort_values(by = 'y_score', ascending = False, inplace = True)# 自定义分位点cuts = np.arange(0.1,1,0.1)# 计算各分位点对应的Score值index = len(target_data.y_score)*cutsscores = np.array(target_data.y_score)[index.astype('int')]# 根据不同的Score值，计算Sensitivity和SpecificitySensitivity = []Specificity = []for score in scores:# 正例覆盖样本数量与实际正例样本量positive_recall = target_data.loc[(target_data.y_test == positive_flag) & (target_data.y_score>score),:].shape[0]positive = sum(target_data.y_test == positive_flag)# 负例覆盖样本数量与实际负例样本量negative_recall = target_data.loc[(target_data.y_test != positive_flag) & (target_data.y_score<=score),:].shape[0]negative = sum(target_data.y_test != positive_flag)Sensitivity.append(positive_recall/positive)Specificity.append(negative_recall/negative)# 构建绘图数据plot_data = pd.DataFrame({'cuts':cuts,'y1':1-np.array(Specificity),'y2':np.array(Sensitivity), 'ks':np.array(Sensitivity)-(1-np.array(Specificity))})# 寻找Sensitivity和1-Specificity之差的最大值索引max_ks_index = np.argmax(plot_data.ks)plt.plot([0]+cuts.tolist()+[1], [0]+plot_data.y1.tolist()+[1], label = '1-Specificity')plt.plot([0]+cuts.tolist()+[1], [0]+plot_data.y2.tolist()+[1], label = 'Sensitivity')# 添加参考线plt.vlines(plot_data.cuts[max_ks_index], ymin = plot_data.y1[max_ks_index], ymax = plot_data.y2[max_ks_index], linestyles = '--')# 添加文本信息plt.text(x = plot_data.cuts[max_ks_index]+0.01,y = plot_data.y1[max_ks_index]+plot_data.ks[max_ks_index]/2,s = 'KS= %.2f' %plot_data.ks[max_ks_index])# 显示图例plt.legend()# 显示图形plt.show()# 导入虚拟数据
virtual_data = pd.read_excel(r'virtual_data.xlsx')
# 应用自定义函数绘制k-s曲线
plot_ks(y_test = virtual_data.Class, y_score = virtual_data.Score,positive_flag = 'P')