A High-Level Architecture of a Metadata-based Ontology Matching Framework

One of the pre-requisites for the realization of the Semantic Web vision are matching techniques which are capable of handling the open, dynamic and heterogeneous nature of the semantic data in a feasible way. Currently this issue is not being optimally resolved; the majority of existing approaches to ontology matching are (implicitly) restricted to processing particular classes of ontologies and thus unable to guarantee a predictable result quality on arbitrary inputs. Accounting for the empirical findings of two case studies in ontology engineering, we argue that a possible solution to cope with this situation is to design a matching strategy which strives for an optimization of the matching process whilst being aware of the inherent dependencies between algorithms and the types of ontologies these are able to process successfully. We introduce a matching framework that, given a set of ontologies to be matched described by ontology metadata, takes into account the capabilities of existing matching algorithms (matcher metadata) and suggests, by using a set of rules, appropriate ones.