基于分层部分核实数据对二项比例的齐性检验  

Homogeneity Test for Binomial Proportions Based on Stratified Partially Validated Data

作　　者：李天骄 LI Tian-jiao(School of Science,Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]重庆理工大学理学院,重庆400054

出　　处：《重庆工商大学学报（自然科学版）》2020年第6期39-47,共9页Journal of Chongqing Technology and Business University:Natural Science Edition

基　　金：重庆市基础研究与前沿探索项目(CSTC2018JCYJAX0241);重庆理工大学研究生创新项目(YCX20192082).

摘　　要：针对分层情况下的疾病流行率问题,考虑了金标准存在时,在各层的敏感度和特异度不同时二项比例(疾病流行率)的齐性检验;提出了基于渐近检验过程的7种统计量和基于Bootstrap重抽样检验过程的4种统计量,并通过蒙特卡罗模拟研究来比较了各种检验的犯第一类错误的概率和检验功效;研究表明:score统计量、似然比统计量和4种基于bootstrap重抽样的检验统计量具有良好的统计性质,推荐用于实际中;最后,实际数据进一步验证了方法的有效性。Aiming at the problem of disease prevalence in stratified situation,the homogeneity test of binomial proportion(disease prevalence)is considered when the sensitivity and specificity of each strata are different in the presence of gold standard.Seven kinds of statistics based on the asymptotic test process and four kinds of statistics based on the bootstrap resampling test process are proposed,and the empirical type I error and the empirical power in various tests are compared by Monte Carlo simulation.The research shows that score statistics,likelihood ratio statistics and four kinds of test statistics based on bootstrap resampling have good statistical properties and are recommended to be used in the practice.Finally,a real data is further used to verify the proposed methods.

关 键 词：金标准 齐性检验 Bootstrap重抽样 分层部分核实数据 

分 类 号：O212.1[理学—概率论与数理统计]