Mapping ontology relations: An approach based on best approximations

Abstract

Relation mappings are important for interoperations between heterogeneous ontologies. Several current methods employ string similarity matching or heuristic rules to find them, but often produce low quality results. We propose a novel approach based on best approximations. The core idea is to find least upper bounds and greatest lower bounds of a relation, and then use them to get upper and lower approximations of the relation. These approximations are the relation mappings between ontologies. To discovery the best mappings, we extend the definition of least upper(/greatest lower) bounds as multielement least upper(/greatest lower) bounds, that not only containing separate relations, but also disjunctions or conjunctions of relations and the related concepts. The simplified multielement bounds are also defined to avoid redundancy. An effective algorithm for finding the relation mappings is proposed. © Springer-Verlag Berlin Heidelberg 2006.