Semantic Mapping between RDBMS and Domain Ontology

In general, although most data existing in the WWW are modeled as some type of relational database, characteristically even information in the same domain is often structured in different ways depending on the individual manager. In order to materialize a semantic Web, a method to efficiently map these different kinds of data is needed. This paper presents an algorithm for the matching between a relational database and domain ontology. Existing studies on the matching between a relational database and domain ontology were performed by extracting local ontology from a relational database. However the problem of losing domain information remains as a correlation with domain ontology is not used in the process of extracting local ontology. As a solution to this problem, we attempted to prevent the information loss through the measurement of similarity between the instances of the relational database and domain ontology, and increased the efficiency of the matching process by using information on the relation between tables in the relational database and also the relation between classes in ontology.