Title: Birnbaum-Saunders linear mixed-effects models with censored data: Bayesian MCMC implementation
Authors: Filidor Vilca - University of Campinas - Unicamp, Brazil (Brazil) [presenting]
Abstract: The use of mixed effects models where responses are clustered around some random effects is usual in analysis of correlated data. The focus is on the Bayesian inference for Birnbaum-Saunders linear mixed models for censored data, which is inspired in previous work from a frequentist viewpoint. Specifically, the use of the Markov chain Monte Carlo method is explored to develop the Bayesian analysis, by using an acceleration convergence procedure. This approach provides an alternative to that developed under frequentist viewpoint that depends on the approximated likelihood function. Bayesian mechanisms for parameter estimation, residual analysis and influence diagnostics are developed. In order to examine the usefulness of this approach, we perform simulation studies taking in account the MCMC algorithm and its modified version. Also, an analysis to real dataset is considered to illustrate the proposed approach.