Building Topic Maps from Relational Databases

Abstract

Relational Databases (RDB) have been traditionally widely used as the backend database for information systems. Considering that RDBs contain valuable data, the challenge is to find out how to improve accessing and sharing knowledge that resides in databases. The use of topic maps is one solution for representing that knowledge. However, manual development of topic maps is a difficult, time consuming and subjective task if there is not a common guideline. The existing topic maps building approaches convert RDBs without considering the knowledge residing in the database.

This paper proposes an automatic approach that considers the background knowledge in the building process of topic maps. The proposed model was implemented and applied on a benchmark of RDBs. The resulted topic maps were validated syntactically using the Ontopia Vizigator tool and validated semantically through the inference of information using the Tolog query language. The results found in our experiments are encouraging.