ON BAYESIAN ANALYSIS OF RIGHT CENSORED WEIBULL DISTRIBUTION USING APPROXIMATE METHODS

  • Navid Feroze Department of Statistics, Riphah International University, Islamabad, Pakistan
  • Muhammad Aslam Department of Statistics, Riphah International University, Islamabad, Pakistan
  • Mariya Raftab Department of Agriculture, Abbatabad, Pakistan
  • Bilal Ahmed Abbasi Department of Management Sciences, The University of Azad Jammu and Kashmir
Keywords: Quadrature Method, Gibbs Sampler, Importance Sampling (IS), Lindley’s Approximation, Tierney and Kadane's Approximation (TKA), Posterior Distributions, Loss Functions

Abstract

In this paper, we have discussed the estimation for parameters of Weibull model under right censored samples. We have assumed two priors and two loss functions for the posterior estimation. As the Bayes estimators from the concerned posterior distributions do not exist in the explicit form, we have considered Quadrature method (QM), Gibbs sampler (GS), importance sampling (IS), Lindley’s approximation (LA) and Tierney and Kadane's approximation (TKA) to obtain the numerical solutions for the Bayes estimators. The performance of the different estimators has been compared using simulated results along with real example. The findings of the study suggest that estimators based on IS and TKA are superior in performance with certain conditions.

 

Downloads

Download data is not yet available.
Published
2018-12-10
Section
Articles