DISCORD IN POSSIBILITY THEORY

Abstract
A function referred to as discord was recently introduced as a justifiable measure of one type of uncertainty in the Dempster-Shafer theory, uncertainty that results from conflicts among evidential claims. In this paper, we investigate mathematical properties of this function within the domain of possibility theory. We show, in particular, that the largest value of discord in possibility theory is about 0.892. This result implies that discord is negligible (in comparison with nonspecificity) in many practical applications of possibility theory, especially when large bodies of evidence are involved.

This publication has 7 references indexed in Scilit: