Exclusive Hypothesis Testing for Cox's Proportional Hazards Model  

作　　者：WU Qiang TONG Xingwei DUAN Xiaogang 

机构地区：[1]School of Statistics,Beijing Normal University,Beijing 100875,China

出　　处：《Journal of Systems Science & Complexity》2024年第5期2157-2172,共16页系统科学与复杂性学报（英文版）

基　　金：supported by the National Natural Science Foundation of China under Grant Nos.11971064,12371262,and 12171374。

摘　　要：Exclusive hypothesis testing is a new and special class of hypothesis testing.This kind of testing can be applied in survival analysis to understand the association between genomics information and clinical information about the survival time.Besides,it is well known that Cox's proportional hazards model is the most commonly used model for regression analysis of failure time.In this paper,the authors consider doing the exclusive hypothesis testing for Cox's proportional hazards model with right-censored data.The authors propose the comprehensive test statistics to make decision,and show that the corresponding decision rule can control the asymptotic TypeⅠerrors and have good powers in theory.The numerical studies indicate that the proposed approach works well for practical situations and it is applied to a set of real data arising from Rotterdam Breast Cancer Data study that motivated this study.

关 键 词：Comprehensive test statistics Cox’s proportional hazards model exclusive hypothesis testing right-censored data Type I error 

分 类 号：O212.1[理学—概率论与数理统计]