Cluster-based Specification Techniques in Dempster-Shafer Theory

TL;DR

This paper introduces a clustering approach within Dempster-Shafer theory to better handle uncertain and weakly specified evidence by partitioning evidences into subsets representing distinct events, including probabilistic assessments of evidence membership.

Contribution

It extends previous work by not only identifying the most plausible subset for each evidence but also calculating the plausibility of belonging to all subsets and estimating the number of subsets probabilistically.

Findings

Effective partitioning of evidence into event-representing subsets.

Calculation of plausibility for evidence belonging to multiple subsets.

Probabilistic estimation of the number of evidence subsets.

Abstract

When reasoning with uncertainty there are many situations where evidences are not only uncertain but their propositions may also be weakly specified in the sense that it may not be certain to which event a proposition is referring. It is then crucial not to combine such evidences in the mistaken belief that they are referring to the same event. This situation would become manageable if the evidences could be clustered into subsets representing events that should be handled separately. In an earlier article we established within Dempster-Shafer theory a criterion function called the metaconflict function. With this criterion we can partition a set of evidences into subsets. Each subset representing a separate event. In this article we will not only find the most plausible subset for each piece of evidence, we will also find the plausibility for every subset that the evidence belongs to the subset. Also, when the number of subsets are uncertain we aim to find a posterior probability distribution regarding the number of subsets.