Bayes Estimators of the Scale Parameter of an Inverse Weibull Distribution under two different Loss Functions

Abstract

In this paper we obtain Bayesian estimators of the scale parameter of the inverse Weibull distribution (IWD).We derive those estimators under two different loss functions: the quasisquared error loss function and the nonlinear exponential loss function (NLINEX). Two priors are considered for finding the estimators: a class of natural–conjugate informative prior, namely; the exponential prior information and inverted-Levy prior information. Based on a Monte Carlo simulation study, the performance of those estimators is compared. The comparison criteria, the mean square errors (MSE) are computed and presented in tables. Comparison results show that MLE was the best followed by Bayes estimators based on the inverse Levy prior under NLINEX loss function which was preferable among the others.