Ontology-driven Resolution of Semantic Heterogeneities in GDB Conceptual Schemas

As the geographic information system (GIS) community grows, more and more people needs to share geographic information. As modern GIS data is stored in geographic databases their conceptual schemas have to be, at least, interoperable. Due to the fact that the databases are designed by many different people from different countries, using different languages and maybe with different definitions for the same phenomenon there is a high probability that the conceptual schemas have semantic heterogeneities between them. In order to handle these heterogeneities this paper suggests the use of ontologies as mediators to the semantic integration. A software architecture is proposed, which handles also syntactic heterogeneities using a standard language, the GML beyond the semantic ones. A ontology that represents a subset of the geographic reality was created and a similarity matching algorithm was developed to process schemas against it. The mathematical methods of similarity measurement modeling concepts have been tuned for a set of real GDB conceptual schemas.