Characterizing Spatial Databases via their Derivation: A Complement to Content Ontologies
Developing an ontology that succinctly describes the contents of a spatial database is a very difficult undertaking. Yet most current efforts to develop spatial ontologies remain focused on describing content. Ontologies describing other aspects of spatial databases may prove to be much easier to develop and nearly as useful as content ontologies, and yet these alternative ontologies have received little attention from the research community. This paper explores one such alternative, specifically, an ontology that describes how a spatial database may have been derived. Derivation ontologies are shown to be highly complementary to content ontologies, and in some cases can perform nearly identical tasks. It is also shown that derivation ontologies are much more straightforward to develop than are content ontologies. Finally, we present a genetic programming (GP)-based approach to automatically developing derivation ontologies for existing databases. It is concluded that while derivation ontologies cannot replace content ontologies, they are a useful and practical complement that offer their own unique set of strengths to the problem of semantically characterizing spatial data.
Document Type: Research Article
Affiliations: Remote Sensing/GIS Program Colorado State University
Publication date: June 1, 2007