User-Driven Ontology Learning from Structured Data

The automatic acquisition of models to represent existing domain knowledge is a key step to further develop domain driven data mining. Ontology Learning has been mostly focused on unstructured data sources, as text, leaving structured data almost ignored. This is probably due to the existence of a model behind that kind of data, that without being an ontology, reveals some data semantics. This paper extends the work by Borgida [1], giving to the user the possibility to choose the level of detail of a domain ontology learnt from a relational database. Beside the full exploration of relational model premises, we apply association rules mining to discover basic axioms, which describe the hidden assertions underlying the domain.