Abstract: It is widely recognized that conditional inference is often as efficient as Bayesian inference based on a non-informative prior. However, it is generally less efficient than Bayesian inference that incorporates an informative prior distribution. Accordingly, the primary objective of this study is to derive conditional point estimates for the parameters of the Gumbel distribution using pivotal quantities, based on generalized progressively hybrid-censored data. These estimates are compared with Bayesian estimates obtained under various loss functions through Monte Carlo simulations. The simulation results indicate that conditional inference is highly efficient and frequently yields better estimates than its Bayesian counterparts. Finally, a real dataset pertaining to a weather phenomenon is analyzed to demonstrate the effectiveness of the proposed methods and to confirm the suitability of the model for such data.

Keywords: Bayes inference, Conditional inference, Informative prior, Kernel prior

Pages: 106-117

WSEAS Transactions on