On the Least Favorable Configuration of a Selection Procedure Based on Ranks.

Abstract

This document considers two types of statistics based on the sums of combined (Wilcoxon type) ranks and vector (Friedman type) ranks. Underlying populations are supposed to belong to the location or scale parameter family of distributions. Two approaches - subset selection and indifference zone - of ranking and selection procedures based on these statistics are considered in an asymptotic framework for selecting the population with the largest parameter value. The least favorable configurations of parameters are discussed by computing the exact moments of these statistics and introducing an assumption of order relation between the gaps of parameters. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1985
Accession Number
ADA159154

Entities

People

  • Sumedha Gupta
  • T. Matsui

Organizations

  • Purdue University

Tags

Communities of Interest

  • C4I
  • Human Systems

DTIC Thesaurus Topics

  • Analysis Of Variance
  • Computing-Related Activities
  • Contracts
  • Covariance
  • Data Science
  • Distribution Functions
  • Information Science
  • Interdisciplinary Science
  • Military Research
  • Observation
  • Probability
  • Statistical Inference
  • Statistics
  • United States
  • United States Government
  • Universities

Fields of Study

  • Mathematics

Readers

  • Calculus or Mathematical Analysis
  • Statistical inference.