Likelihood and Bayesian estimations for step-stresslife test model under Type-I censoring

Abstract

This paper discusses likelihood and Bayesian estimations for partiallyaccelerated step-stress life test model under Type-I censoring assuming Pareto distribution of the second kind. The posterior means andposterior variances are obtained under the squared error loss functionusing Lindley's approximation procedure. It has been observed thatLindley's method usually provides posterior variances and mean squareerrors smaller than those of the maximum likelihood estimators. Furthermore, the highest posterior density credible intervals of the modelparameters based on Gibbs sampling technique are computed. For illustration, simulation studies and an illustrative example based on areal data set are provided