Ontology representation and reasoning : a conceptual level approach
Ontologies play a key role in many areas of Computing Science, such as Information Retrieval, Knowledge Management, and Knowledge Engineering. However, Ontology development and maintenance is a challenging task that is currently not very well supported by software tools. Most existing ontology editors cannot provide the kind of automated reasoning support that is required for the verification and for the validation of ontologies. More concretely, such an automated support should (i) check the ontology consistency and (ii) suggest possible enhancements to the ontology taxonomy. Description Logic engines compute the kind of inferences that are useful for an automated ontology verification and validation but are not suitable for all ontology representation languages. More concretely, the semantics of Description Logics is based on the Open World Assumption, whereas the semantics of some ontology representation languages is based on the Closed World Assumption. Furthermore, the knowledge model of Description Logics is derived from Frame-based knowledge representation. Therefore Description Logics lack some modelling primitives necessary to express knowledge that can be represented with conceptual modelling languages. On the other hand, conceptual modelling languages (i) do not have the same expressive power as Frame-based ontology languages and (ii) no reasoners are available for automated reasoning with these languages. Hence, this thesis introduces the Conceptual Knowledge Modelling Language (CKML) and proposes an approach for the verification and validation of CKML ontologies. Rather than developing a special-purpose reasoning algorithm for CKML, we investigate how Description Logic engines can be used for this task. This approach can also be applied to a language that describes database schemas specified with the Enhanced Entity-Relationship model.