On Some Nonparametric Selection Procedures
Abstract
In many practical situations, the experimenter often faces the problem of comparing several populations, treatments in clinical trials or processes. The selection and ranking method of ranking and selection provides the useful techniques for solving such problems. There have been two main approaches to selection and ranking problems, the indifference zone approach and the subset selection approach. In the indifference zone approach a single population is chosen and is guaranteed to be the best (worst) with probability at least equal to P*. However, in this formulation it is assumed that the best population is sufficiently apart from the remaining k - 1 populations. In the subset selection approach no such restriction on the parameter space is assumed. A random size subset of k populations is chosen which is guaranteed to contain the best (worst) population with probability at least equal to P*. In this approach the data or the outcome of the experiment is used to decide on how many populations to select. Nonparametric statistics; Equations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1990
- Accession Number
- ADA223831
Entities
People
- Sayaji N. Hande
- Shanti Gupta
Organizations
- Purdue University