The Extended Jeffrey Prior Based on Generalized Fisher Information

Authors

  • Muhammad Fajar Badan Pusat Statistik-Statistics Indonesia, Indonesia

DOI:

https://doi.org/10.33603/yapyh614

Keywords:

Jeffrey prior, fisher information, Bayesian, Posterior

Abstract

The Jeffrey prior is noninformative prior formed from the Fisher information. However, nowadays, Fisher information is extended to generalized fisher information, which has various forms of specifications. Therefore, this paper aims to establish a theoretical framework of the extended Jeffrey prior based on generalized Fisher information as a first step of investigation for future research. The findings of this study are that the form of the extended Jeffrey prior varies depending on the type of generalized fisher information used. However, the posterior formed using the extended Jeffrey prior is not closed-formed and requires a more complex estimation algorithm than the standard algorithm.

Downloads

Published

2024-05-05

Issue

Section

Artikel

Citation Check