Comprehensive Comparative Study of Non-Bayesian and Bayesian Estimators for the DUSTopp-Leone Burr-Hatke-Exponential Model

Abstract

This study investigates both non-Bayesian and Bayesian methods for estimating the parameters of the DUSTLBHE model. Specifically, the parameters are estimated using nineteen non-Bayesian methods as well as Bayesian method under three distinct loss functions. To assess the efficiency of these estimators, an extensive Monte-Carlo simulation study is carried out complemented by real dataset applications from the medical field. The performance of the estimation methods is evaluated using statistical measures such as the root mean squared error (RMSE) and the Kolmogorov-Smirnov (KS) test. Simulation findings suggest that non-Bayesian estimators most notably maximum likelihood (MLE) and Maximum product spacing (MPS) estimators consistently achieve superior accuracy compared to Bayesian methods when estimating the parameters, with performance steadily improving as the sample size increases. The real-datasets analyses show that the Anderson-Darling (ANDA) and percentile (PE) estimators produce the most accurate parameter estimates for the DUSTLBHE parameters.