Some Results on Subset Selection Problems.
Abstract
The main purpose of this paper is to propose and study the subset selection approach for some new problems. Chapter I deals with some subset selection procedures for Poisson populations. A procedure of the type discussed by Seal is proposed and compared with the main proposed procedure which is an unconditional rule. Some selection procedure for populations better than a standard are also investigated. In Chapter II, a subset selection procedure based on the sample median for selecting the largest of the k location parameters of double exponential (Laplace) distributions is studied. For this distribution the problem of selection for the scale parameters is also investigated. A test of homogeneity is proposed which is based on the range of sample medians. An indifference zone approach to the problem of selecting the populations with the t-largest unknown means is also studied. Chapter III discusses some classification rules for k univariate normal populations using the subset selection approach. The classification problem is studied in terms of (i) the mean (ii) the variance and (iii) the reciprocal of the coefficient of variation. Chapter IV deals with a conditional and an unconditional procedure for selecting a subset which contains the negative binomial population with the smallest unknown probability of a success. Selection of populations better than a standard is also investigated. An application of the procedure to reliability theory is described. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1976
- Accession Number
- ADA033362
Entities
People
- Yoon-kwai Leong
Organizations
- Purdue University