Author(s)

S. Shi, M.E.C. Hull & D.A. Bell

Abstract

How to combine and update evidence when conducting reasoning with un- certainty is a key topic in knowledge-based systems. In our previous work we have observed that the insufficient reasoning principle of probability the- ory is a popular concept which dominates the suggested rules for updating beliefs and combining evidence when handling uncertainty. In this paper, based on this observation we consider a new rule for combining evidence in probabilistic reasoning. This rule corresponds to Dempster's rule of combi- nation in the Dempster-Shafer theory of evidence and contrasts to Bayes' rule of conditionalization and Jeffrey's generalization of conditionalization which are updating rules. The result is that two rules are available for probabilistic reasoning: one is the Bayes' rule of c

Keywords