Handling topical metadata regarding the validity and completeness of multiple-source information: A possibilistic approach

Abstract

We study the problem of aggregating metadata about the validity and/or completeness, with respect to given topics, of information provided by multiple sources. For a given topic, the validity level reflects the certainty that the information stored is true. The completeness level of a source on a given topic reflects the certainty that a piece of information that is not stored is false. We propose a modeling based on possibility theory which allows the fusion of such multi-source information into a graded belief base.