Classical and Bayes-P Subset Selection Procedures for Double Exponential Populations
Abstract
The exact distribution of the sample mean from a double EXPONENTIAL(Laplace) model is derived. A classical subset selection procedure based on the sample mean for selecting the population associated with the largest location parameter of k double exponential(Laplace) distributions is studied. For the case when a non-informative prior is introduced into the problem, the relation between the classical Maximum-Type Procedure Rule Rmax and the so-called Bayes-P* subset selection procedure rule is studied. An improved bound for the guarantee probability of a correct selection for the classical subset selection rule Rmax that relates the rule Rmax to the selected subset size (notice that the subset selection rule Rmax may select all the populations) is studied and some improved rules of the type Rmax are provided. Exponential functions; Laplace transformations; Bayes theorem.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1990
- Accession Number
- ADA223830
Entities
People
- Shanti Gupta
- Yuning Liao
Organizations
- Purdue University