A DL-based Approach for Detecting Semantic Relations in Geo- Ontology Matching

Ontology matching produces mapping relations between elements of two ontologies and it is a basic problem in geographical information integration. Currently, most existing studies rely on all kinds of semantic similarity between the semantic entities to measure ontology mapping relations. However, these measures are not sufficient due to only detecting equivalence relation of compared ontologies, these results also fail to consider potential restriction conflicts, syntax analysis and semantics reasoning techniques. In addition, evaluation studies have shown that general precision and recall can merely cover a fraction of the semantic matching ontology involved, and error-prone problems inevitably exist. In this study, we presented a DL-based method for detecting semantic relations in geo-ontology matching. First, we use Protege4.1 to build geo-ontologies based mainly on the fundamental geographic information data related to land use from the second national land survey of China. Second, we improve an existing matching algorithm of ontologies and identified the lexical matching relation by using the professional dictionary. Finally, we exploit the reason process for detecting semantic relations in geo-ontology in Eclipse environment and implement inconsistency check. Experiments show that the method is feasible and effective in the mapping of geo-ontologies.