Estimating Quantiles of Two Exponential Populations under Ordered Location Using Censored Samples

Abstract

The problem of component wise estimation of quantiles of two shifted exponential populations has been considered under type-II censored samples when the location parameters assume certain ordering. When there is no order restriction on the location parameters, estimators like maximum likelihood estimator (MLE), modified maximum likelihood estimator (MMLE), uniformly minimum variance unbiased estimator (UMVUE) and best affine equivariant estimator (BAEE) have been found. Incorporating the ordered restriction on the location parameters, isotonic estimators of the BAEE and the mixed estimators have been obtained. Further, using prior information of ordered location parameters, certain Bayes estimators have been obtained. All the proposed estimators have been compared using Monte-Carlo simulation technique. Finally conclusions have been made regarding the use of the estimators.