MAMA: A Novel Approach to Ontology Mapping

Ontology is a key factor for enabling interoperability across heterogeneous systems, services and users. One of the most challenging tasks in its use is ontology construction. However, experts usually represent the same knowledge domain by the use of different ontologies that can differ in structure, concepts, relations and attributes. Therefore, ontology could lose its main feature: allowing the semantic interoperability among actors working on the same knowledge domain. An effective solution to this problem is the introduction of methods for finding mapping among the various components of ontologies. In this paper, a novel approach to the ontology mapping is proposed. The proposed approach, named MAMA, investigates the possibility of finding overlaps among different ontologies that describe the same knowledge domain through the combined use of syntactic, semantic and topological similarity indexes. The indexes are combined to define the degree of similarity between the various components of ontologies by introducing a combining rule. This rule is adaptive and automatically emphasizes the contributions of those indexes that provide better results in certain operative conditions. The proposed approach has been tested on standard datasets and the obtained results are very promising. Moreover, we are currently exploring the application of the ontology mapping approach to software component reuse.