The generalized entropy measure to the design and comparison of regression experiment in a Bayesian context

Abstract

Taneja [14] studied a unified (r, s)-entropy that includes as a particular case some of the known entropies. Based on this unified (r, s)-entropy, Pardo et al. [8] defined the average amount of information provided by an experiment X over the unknown parameter θ with prior knowledge p(θ). By using average amount of information in unified form, we compare experiments based on the Bayesian approach. Some connections with the criterion of Blackwell and Lehmann are also made. In this paper, an application of generalized entropy measures to the design and comparison of linear regression experiment is presented.