Knowledge Transformation using a Hypergraph Data Model

In the Semantic Web, knowledge integration is frequently performed between heterogeneous knowledge bases. Such knowledge integration often requires the schema expressed in one knowledge modelling language be translated into an equivalent schema in another knowledge modelling language. This paper defines how schemas expressed in OWL-DL (the Web Ontology Language using Description Logic) can be translated into equivalent schemas in the Hypergraph Data Model (HDM). The HDM is used in the AutoMed data integration (DI) system. It allows constraints found in data modelling languages to be represented by a small set of primitive constraint operators. By mapping into the AutoMed HDM language, we are then able to further map the OWL-DL schemas into any of the existing modelling languages supported by AutoMed. We show how previously defined transformation rules between relational and HDM schemas, and our newly defined rules between OWL-DL and HDM schemas, can be composed to give a bidirectional mapping between OWL-DL and relational schemas through the use of the both-as-view approach in AutoMed.