期刊名称：Pakistan Journal of Statistics and Operation Research
印刷版ISSN：2220-5810
出版年度：2014
卷号：10
期号：1
页码：17-39
DOI：10.1234/pjsor.v10i1.649
语种：English
出版社：College of Statistical and Actuarial Sciences
摘要：The paper is concerned with the preference of prior for the Bayesian analysis of the shape parameter of the mixture of Burr type X distribution using the censored data. We modeled the heterogeneous population using two components mixture of the Burr type X distribution. A comprehensive simulation scheme, through probabilistic mixing, has been followed to highlight the properties and behavior of the estimates in terms of sample size, corresponding risks and the proportion of the component of the mixture. The Bayes estimators of the parameters have been evaluated under the assumption of informative and non-informative priors using symmetric and asymmetric loss functions. The model selection criterion for the preference of the prior has been introduced. The hazard rate function of the mixture distribution has been discussed. The Bayes estimates under exponential prior and precautionary loss function exhibit the minimum posterior risks with some exceptions.
