Comparison of improved class of priors for the analysis of the Burr type VII model under doubly censored samples

Authors : Navid FEROZE, Muhammad ASLAM

Pages : 1509-1533

Doi:10.15672/hujms.989479

View : 23 | Download : 13

Publication Date : 2021-10-15

Article Type : Research Paper

Abstract :In recent years, the finite mixture lifetime models have frequently been used in chemical, physical, social science, biological and other fields due to their methodological development and practical applications. The Bayesian analysis of the mixture models has also developed a significant interest among the statisticians especially in the last decade. However, the most of these contributions are limited to the Bayes estimation for the parameters of lifetime models under singly type I censoring. This paper discusses the Bayesian estimation for the two-component mixture of the lifetime models under doubly censored samples with a particular case for the Burr type VII model. A class of improved priors has been proposed for the posterior estimation. The likelihood function, for doubly censored samples using two-component mixture of life time distributions, has been introduced. The hazard rate function for the mixture model has been compared for different parametric values. The performance of various estimators has been compared under a simulation study along with a real life example.
Keywords : Mixture models, posterior distributions, loss functions, credible intervals, posterior predictive intervals