Aggregation of Possibilities and Certainty Factors with Dempster's Rule of Combination

Abstract

Uncertainty combination is an important part of information processing by humans, but there are only a few theories that contain a standard way of uncertainty combination. Recently, some new combination rules dealing with binary hypothesis have been proposed in the forms of Possibility and Certainty Factors, where the information source model is defined rigorously and CFs are interpreted as another expression of possibility. The paper proposes additional two combination rules in the same framework as the above using Dempster's rule in Dempster-Shafer theory of Evidence. It also discusses the mathematical properties of the combination rules as well as properties for applications using numerical examples.