Approximations of concept based on multielement bounds

Abstract

Ontology-based information system can increase the precise and recall of information retrieval and filtering on the Web but faces the problem of Ontology heterogeneity. The approximate information filtering approach rewrites queries with their approximations to suit specific ontology-based information sources. The core of the approach is to find approximations of concepts. The current methods find least upper bounds and greatest lower bounds of a concept and then use them to get upper and lower approximations of the concept. However, they consider the bounds only containing separate concepts, so the quality of approximations is not very acceptable. This paper defines multielement least upper bounds and multielement greatest lower bounds that contain not only separate concepts but also disjunctions or conjunctions of concepts. It can be proved that the approximations based on them can be better approximations of concepts than the current methods. © Springer-Verlag Berlin Heidelberg 2005.