On the Least Favorable Configuration of a Selection Procedure Based on Ranks.
Abstract
The problem of selection of the population with the largest parameter is considered using the subset selection as well as the indifference zone approach for distributions that belong to a location or a scale parameter family. The procedures are based on the sums of combined (Wilcoxon type) ranks and vector (Friedman type) ranks. The least favorable configurations are obtained in an asymptotic framework under certain order relations between the gaps of parameters. The asymptotic theory is based on exact moments of the rank sum statistics. Keywords: Rank order Statistics; Computations; Population(Mathematics); Value.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1987
- Accession Number
- ADA178454
Entities
People
- Shanti Gupta
- Takashi Matsui
Organizations
- Purdue University