Product Spacings as an Alternative to Likelihood for Bayesian Inferences

Abstract

The paper aims to study the consequences of using product of spacings (PS) as an alternative to the traditional likelihood under Bayesian set up. For this purpose we have considered the problem of point estimation of the parameter of exponential distribution. We have also obtained the asymptotic and HPD confidence intervals of the parameter. The proposed estimates have been compared with those based on usual likelihood on the basis of simulated samples from exponential distribution.