Research On Multiple Testing And The Estimate Of Its Parameter

As one of the main research areas in statistical inference,hypothesis test-ing has more and more extensive application in a variety of horizons such as the research and development of clinical medicine, the promote of com-modity and genomics studies.The Single hypothesis testing theory has been improved. But with the emergence of mega high-dimensional data, multiple testing problems,which was quite popular in the 1950s,attracts many statisti-cians’attentions,especially the application of FWER and FDR in the analysis of the differential expression of microarray data.In this paper, we focus on multiple hypothesis testing.In Chapter 1, we first introduce the basic concepts of hypothesis testing and the background and significance of the research on multiple testing. In Chapter 2, we describe several testing procedures and different, methods.We also compare their dif-ferences.In Chapter 3, we consider the π0, which is an important parameter in the improvements of testing proceduers.Many statisticians have proposed many estimators. We mainly discuss several of them and propose EM method as a direct way to estimate π0 and combine it with difference met hod. In the last Chapter, we compare the methods above by simulation and analyse their characteristics.