Chi-Square Goodness of Fit Test | Hosmer–Lemeshow test

最新推荐文章于 2025-05-01 17:35:42 发布

条件漫步

最新推荐文章于 2025-05-01 17:35:42 发布

阅读量318

点赞数

文章标签： python Hosmer-Lemeshow

本文链接：https://blog.youkuaiyun.com/chenhepg/article/details/129650558

版权

创建于：2023.3.19
修改于：2023.3.19

文章目录

- 1、Chi-Square Goodness of Fit Test
- 2、Hosmer–Lemeshow test 的理解以及Python代码

1、Chi-Square Goodness of Fit Test

在这里插入图片描述

Chi-Square Goodness of Fit Test | Formula, Guide & Examples

2、Hosmer–Lemeshow test 的理解以及Python代码

在这里插入图片描述

from scipy.stats import chi2
def Hosmer_Lemeshow_test(data, Q=10):
    '''
    data: dataframe format, with ground_truth label name is y,
                                 prediction value column name is y_hat
    '''
    data = data.sort_values('y_hat')
    data['Q_group'] = pd.qcut(data['y_hat'], Q)
    
    y_p = data['y'].groupby(data.Q_group).sum()
    y_total = data['y'].groupby(data.Q_group).count()
    y_n = y_total - y_p
    
    y_hat_p = data['y_hat'].groupby(data.Q_group).sum()
    y_hat_total = data['y_hat'].groupby(data.Q_group).count()
    y_hat_n = y_hat_total - y_hat_p
    
    hltest = (((y_p - y_hat_p)**2 / y_hat_p) + ((y_n - y_hat_n)**2 / y_hat_n)).sum()
    pval = 1 - chi2.cdf(hltest, Q-2)
    
    print('\n HL-chi2({}): {}, p-value: {}\n'.format(Q-2, hltest, pval))
    return