Test of independence of a class of bivariate distributions based on mutual information

In this paper we studied the entropies and the mutual information of several bivariate distributions that are used in reliability theory. We observed that the mutual information of these distributions are identical and is a function of the dependence parameter only. We proposed a general class of bi...

Full description

Saved in:
Bibliographic Details
Published in:Communications in statistics. Theory and methods 1997-01, Vol.26 (11), p.2803-2816
Main Authors: Ahsanullah, M., Habibullah, Mohamed
Format: Article
Language:English
Subjects:
dependence parameter
entropy
reliability theory
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper we studied the entropies and the mutual information of several bivariate distributions that are used in reliability theory. We observed that the mutual information of these distributions are identical and is a function of the dependence parameter only. We proposed a general class of bivariate distributions whose members have this common mutual information. We developed the maximum likelihood estimator of this dependence parameter and hence the estimator of the mutual information for this class of distributions. We used this estimator to construct an exact test of independence of the random variables and studied its power function
ISSN:0361-0926
1532-415X
DOI:10.1080/03610929708832078