Hybrid approach towards ontology mapping

Abstract

It is common for different ontologies to exist for similar domains, which leads to ontology heterogeneity. This can be addressed by ontology mapping. Issues related to ontology mapping include the lack of efforts to hybridize the various ontology mapping algorithms that are available. Another problem concerns the mapping approaches themselves, i.e. these are either concept-based or instance-based (and not a combination of these). In this paper, we present an approach for ontology mapping. Our approach employs a library of ontology mapping algorithms. A number of these algorithms are selected and linked or hybrided to produce a composite algorithm. Furthermore, we also present a result aggregation engine which can either present results of the individual algorithms separately or fuse the results of the composite algorithm.