Classification Rules for Exponential Populations: Two Parameter Case.
Abstract
This paper extends many of the results for one-parameter exponential distributions, considered by the authors earlier (Proc. Conf. Reliability and Biometry, 1974), to the case of the two parameter exponential distributions. In this paper the classification rules based on the likelihood-ratio criterion for the two-parameter exponential populations have been studied. These classification rules have been extended to many populations and to the situation where observations are censored. In cases where the likelihood-ratio criterion does not lead to appealing classification rules, ad hoc rules are proposed. An additional rule is considered from 'life testing' point of view. In each case, the probability of misclassification is derived exactly where the parameters are known. The rules considered have been shown to possess a consistency property. An alternate approach based on Bayesian considerations is also explored. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1976
- Accession Number
- ADA025181
Entities
People
- Abhinav Gupta
- Asit P. Basu
Organizations
- University of Missouri