Title: Bayesian estimation techniques in certain life-testing models
Authors: Inmaculada Barranco-Chamorro - Fundacion de investigacion de la Universidad de Sevilla (FIUS) (Spain) [presenting]
Abstract: In life-testing and reliability models, mainly due to economic reasons, the experiments must finish before all the units in the sample fail. So it is common to deal with censored data. One of the most important censoring schemes, proposed in recent years, is progressive type-II right censoring in which surviving units can be censored at any time of failure. Under this censoring scheme, bayesian results are given when sampling from the Burr type-XII and the generalized half-logistic distribution. Both models are widely used in life-testing. Expressions are proposed for the Bayes and posterior risks of Bayes estimators under different loss functions. A simulation study is carried out to show the importance of the different features involved in the process of estimation.