Measuring Disagreement among Knowledge Bases

Abstract

When combining beliefs from different sources,<br/>often not only new knowledge but also conflicts arise.<br/>In this paper, we investigate how we can measure the disagreement<br/>among sources. We start our investigation with disagreement measures<br/>that can be induced from inconsistency measures in an automated way.<br/>After discussing some problems with this approach,<br/>we propose a new measure that is inspired by the $\eta$-inconsistency measure.<br/>Roughly speaking, it measures how well we can satisfy all sources simultaneously.<br/>We show that the new measure satisfies desirable properties,<br/>scales well with respect to the number of sources and illustrate its applicability<br/>in inconsistency-tolerant reasoning.