A subset selection procedure for exponential populations under random censoring

Publisher Website
Google Scholar
Abstract
The be k exponential populations. The problem of subset selection for these k populations is formulated in order to accommodate randomly censored observations. A selection procedure is proposed, based on the maximum likelihood estimators of θi i = 1, 2, … , k. It is shown that the is independent of the meanlife ξ of the censoring distribution and that the specification of P∗ - value depends on k. This dependency is one of the distinct features that is inherent to the selection procedure under random censoring. Some desirable properties and the condition under which Gupta's (1963) multiplicative constants can be used in the proposed selection procedure are discussed. In the context of reliability, one can employ the same procedure to select the best population with respect to the hazard rate or system reliability.

This publication has 9 references indexed in Scilit: