一文完全理解模型ks指标含义并画出ks曲线（包含代码和详细解释）「建议收藏」

import numpy as np
import pandas as pd


def ks(df, y_true, y_pre, num=10, good=0, bad=1):
    # 1.将数据从小到大平均分成num组
    df_ks = df.sort_values(y_pre).reset_index(drop=True)
    df_ks['rank'] = np.floor((df_ks.index / len(df_ks) * num) + 1)
    df_ks['set_1'] = 1
    # 2.统计结果
    result_ks = pd.DataFrame()
    result_ks['group_sum'] = df_ks.groupby('rank')['set_1'].sum()
    result_ks['group_min'] = df_ks.groupby('rank')[y_pre].min()
    result_ks['group_max'] = df_ks.groupby('rank')[y_pre].max()
    result_ks['group_mean'] = df_ks.groupby('rank')[y_pre].mean()
    # 3.最后一行添加total汇总数据
    result_ks.loc['total', 'group_sum'] = df_ks['set_1'].sum()
    result_ks.loc['total', 'group_min'] = df_ks[y_pre].min()
    result_ks.loc['total', 'group_max'] = df_ks[y_pre].max()
    result_ks.loc['total', 'group_mean'] = df_ks[y_pre].mean()
    # 4.好用户统计
    result_ks['good_sum'] = df_ks[df_ks[y_true] == good].groupby('rank')['set_1'].sum()
    result_ks.good_sum.replace(np.nan, 0, inplace=True)
    result_ks.loc['total', 'good_sum'] = result_ks['good_sum'].sum()
    result_ks['good_percent'] = result_ks['good_sum'] / result_ks.loc['total', 'good_sum']
    result_ks['good_percent_cum'] = result_ks['good_sum'].cumsum() / result_ks.loc['total', 'good_sum']
    # 5.坏用户统计
    result_ks['bad_sum'] = df_ks[df_ks[y_true] == bad].groupby('rank')['set_1'].sum()
    result_ks.bad_sum.replace(np.nan, 0, inplace=True)
    result_ks.loc['total', 'bad_sum'] = result_ks['bad_sum'].sum()
    result_ks['bad_percent'] = result_ks['bad_sum'] / result_ks.loc['total', 'bad_sum']
    result_ks['bad_percent_cum'] = result_ks['bad_sum'].cumsum() / result_ks.loc['total', 'bad_sum']
    # 6.计算ks值
    result_ks['diff'] = result_ks['bad_percent_cum'] - result_ks['good_percent_cum']
    # 7.更新最后一行total的数据
    result_ks.loc['total', 'bad_percent_cum'] = np.nan
    result_ks.loc['total', 'good_percent_cum'] = np.nan
    result_ks.loc['total', 'diff'] = result_ks['diff'].max()
    
    result_ks = result_ks.reset_index()
    
    return result_ks

import matlibplot.pyplot as plt
import seaborn as sns
sns.set()

def ks_curve(df, num=10):
    # 防止中文乱码
    plt.rcParams['font.sans-serif']=['SimHei']
    plt.rcParams['axes.unicode_minus']=False
	
	ks_value = df['diff'].max()
    # 获取绘制曲线所需要的数据
    x_curve = range(num + 1)
    y_curve1 = [0] + list(df['bad_percent_cum'].values[:-1])
    y_curve2 = [0] + list(df['good_percent_cum'].values[:-1])
    y_curve3 = [0] + list(df['diff'].values[:-1])
    # 获取绘制ks点所需要的数据
    df_ks_max = df[df['diff'] == ks_value]
    x_point = [df_ks_max['rank'].values[0], df_ks_max['rank'].values[0]]
    y_point = [df_ks_max['bad_percent_cum'].values[0], df_ks_max['good_percent_cum'].values[0]]
    # 绘制曲线
    plt.plot(x_curve, y_curve1, label='bad', linewidth=2)
    plt.plot(x_curve, y_curve2, label='good', linewidth=2)
    plt.plot(x_curve, y_curve3, label='diff', linewidth=2)
    # 标记ks
    plt.plot(x_point, y_point, label='ks - {:.2f}'.format(ks_value), color='r', marker='o', markerfacecolor='r', markersize=5)
    plt.scatter(x_point, y_point, color='r')
    plt.legend()
    plt.show()
    
    return ks_value

result_ks = ks(df, 'label', 'score')
ks_curve(result_ks)

from sklearn.metrics import roc_curve

fpr, tpr, thresholds= roc_curve(df.label, df.score)
ks_value = max(abs(fpr-tpr))

# 画图，画出曲线
plt.plot(fpr, label='bad')
plt.plot(tpr, label='good')
plt.plot(abs(fpr-tpr), label='diff')
# 标记ks
x = np.argwhere(abs(fpr-tpr) == ks_value)[0, 0]
plt.plot((x, x), (0, ks_value), label='ks - {:.2f}'.format(ks_value), color='r', marker='o', markerfacecolor='r', markersize=5)
plt.scatter((x, x), (0, ks_value), color='r')
plt.legend()
plt.show()

ks_value = max(abs(fpr-tpr))
# 画图，画出曲线
plt.plot(fpr, tpr)
plt.plot([0,1], [0,1], linestyle='--')
# 标记ks
x = np.argwhere(abs(fpr-tpr) == ks_value)[0, 0]
plt.plot([fpr[x], fpr[x]], [fpr[x], tpr[x]], linewidth=4, color='r')
# plt.scatter((x, x), (0, ks_value), color='r')
plt.xlabel('False positive', fontsize=20)
plt.ylabel('True positive', fontsize=20)
plt.show()