Some Contributions to Gamma-Minimax and Empirical Bayes Selection Procedures.

摘要

The inadequacies of classical tests of homogeneity, in certain situations of interest to the experimenter, have long been known. Ranking and selection procedures which are formulated to overcome these shortcomings have provided realistic methods to handle the problems of comparing k populations. In this thesis, some results on gamma-minimax selection procedures and empirical Bayes selection procedure have been obtained. In Chapter I, a problem of selecting populations close to a control is considered. Our goal is to select those populations which are sufficiently close to the control. Chapter II deals with the problem of selecting the t-best populations. In Chapter III, a problem of selecting populations better than a control is considered.