The Fractional Bayes Factor Approach to the Bayesian Testing of the Weibull Shape Parameter

The techniques for selecting and evaluating prior distributions are studied over recent years which the primary emphasis is on noninformative priors. But, noninformative priors are typically improper so that such priors are defined only up to arbitrary constants which affect the values of Bayes factors. In this paper, we consider the Bayesian hypotheses testing for the Weibull shape parameter based on fractional Bayes factor which is to remove the arbitrariness of improper priors. Also we present a numerical example to further illustrate our results.