Preference of Estimation Approach for Rayleigh Progressive Type II Data

Abstract

This paper compare the performance of the empirical Bayes and generalized maximum likelihood estimation approaches in context of progressively Type II censored data from one parameter Rayleigh distribution. The generalized maximum likelihood and empirical Bayes estimates of scale parameter, reliability function, and failure rate function are compared using risk efficiency criterion. The empirical Bayes estimates are considered with respect to squared error loss function. The wind speed data is presented to illustrate the proposed estimation approaches, and an extensive Monte Carlo simulated study is done to compare the empirical Bayes and Generalized maximum likelihood estimates. The study indicates that the empirical Bayesian approach using squared error loss function is preferable than the generalized maximum likelihood approach for the estimation of reliability performances.

Keywords: Progressively Type II censored samples, generalized maximum likelihood estimation, squared error loss function, empirical Bayes estimation, Risk efficiency, Monte Carlo simulation.