Approaches to ontology integration: Federated SPARQL queries and SWRL

When considering OWL ontologies, a common approach to information integration is mapping equivalency relationships between classes, properties, and individuals in the ontologies. Here, we describe two approaches to solving the problem of automatically matching identical individuals in separate ontologies and merging information about those individuals. The first approach involves a graphical tool for building and executing federated SPARQL queries to integrate information from distributed sources. In such a distributed environment, inference engines will not perform well, and the logical relationships between resources in the ontologies is encoded in the queries themselves, which may be saved and re-executed to build applications that are sensitive to dynamic changes in the data sources. The second approach involves building SWRL rules to identify equivalent individuals between ontologies. Applying the rules will cause new OWL assertions about the individuals to be generated, allowing an application to harness the power of existing OWL inference engines. Finally, we will demonstrate an application of these techniques using currently available technology.