Ontology mapping based on concept classification

Abstract

Ontology mapping is a crucial task for achieving large scale semantic inter-operation of heterogeneous information sources. Conventional mapping solutions mainly focus on the mapping of a pair of ontologies. While such approaches are effective in creating one-to-one ontology mappings, they are less efficient when dealing with the many-to-many ontology mapping scenarios. To cope with the complexity issue faced by pair-wise mapping solutions, we propose a many-to-many ontology mapping approach. By treating ontology mapping as a concept classification problem, we aim to reduce the number of concept comparisons involved in many-to-many ontology mapping. A preliminary evaluation of performance was also carried out, and results show that the proposed mapping solution requires less concept comparison cycles for many-to-many ontology mapping.