Two New Classes of Subset Selection Procedures for Location Parameters

Abstract

The problem of selecting a subset containing the largest of several location parameters of k populations is considered. The proposed classes of selection procedures are based on two-sample U-stalistics. Rank representation of the LJstatistics is given, which facilitates the computation of these statistics. These procedures are shown to satisfy the P^-condition. Strong monotonicity, monotonicity and unbiasedness of the proposed procedures is also established.