Estimating the True Probability of Correct Selection for Location and Scale Parameter Families.
Abstract
The usual formulation of the problem of selecting the best of k populations has been studied from the point of view of designing an experiment that will guarantee a certain probability of making a correct selection (PCS). However, in retrospective studies the sample has already been taken and we wish to estimate the probability of making a correct selection using this data. The present study deals with obtaining a class of upper and lower bounds for the true PCS and its estimator using the methods of majorization. We also study the large sample behavior of the estimator of the PCS, which turns out to be asymptotically normal. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 28, 1976
- Accession Number
- ADA033341
Entities
People
- I. Olkin
- M. Sobel
- Y. L. Tong
Organizations
- Stanford University