Classification Rules for Two Exponential Populations with a Common Location Using Censored Samples

Abstract

The problem of classification into two exponential populations with a common location parameter under the type-II censoring scheme is considered. Estimators improving upon the MLEs and the UMVUEs are used to construct classification rules for classifying an observation as well as a group of observations. Further, a rule-based on the generalized likelihood ratio test, has also been proposed. More importantly, a detailed and in-depth simulation study has been carried out in order to compare the probability of correct classification as well as expected probability of correct classification numerically. Finally, a real life example is presented in order to illustrate the applicability of the proposed classification rules, under type-II censoring scheme.

Keywords

• maximum likelihood estimators
• probability of correct classification
• rules based on generalized likelihood ratio test
• simulation study
• uniformly minimum variance unbiased estimator