Bayes Factor for Testing the Equality of Two Inverse Gaussian Population

• 발행기관: 한국자료분석학회
• 저자명: 김영훈(Yeung-Hoon Kim) 고정환(Jeong Hwan Ko)
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.7 No.4, 1145~1153쪽, 전체 9쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2005.08.30

영문 초록

In this paper, we address Bayesian hypotheses testing for the comparison of inverse Gaussian distributions. In Bayesian testing problem, the conventional Bayes factors can not typically accommodate the use of noninformative priors which are improper and are defined only up to arbitrary constants. To overcome such problem, we consider a Bayesian testing for the equality of two independent inverse Gaussian means using the intrinsic Bayes factors suggested by Berger and Pericchi(1996, 1998). As prior distribution for the parameters, we assumed the noninformative improper priors. The testing results by Bayes factor are compared with those by the classical method via real data analysis.

목차

1. Introduction

2. The Intrinsic Bayes Factor Methodology

3. Bayesian Test for the Equality of Inverse Gaussian Distribution

4. Illustrative Examples

Reference

키워드

Intrinsic Bayes factor Reference prior Inverse Gaussian Minimal training sample Marginal probability density function