A Structure-Based Similarity Spreading Approach for Ontology Matching

Most of the frequently used ontology mapping methods to date are based on linguistic information implied in ontologies. However, same concepts in different ontologies can represent different semantics under the context of different ontologies, so relationships on mapping cannot be solely recognized by applying linguistic information. Discovering and utilizing structural information in ontology is also very important. In this paper, we propose a structure-based similarity spreading method for ontology matching which consists of three steps. We first select centroid concepts from both ontologies using similarities between entities based on their linguistic information. Second, we partition each ontology based on the set of centroid concepts recognized in it using clustering method. Third, we utilize a similarity spreading method to update the similarities between entities from two ontologies and apply a greedy matching method to establish the final mapping results. The experimental results demonstrate that our approach is very effective and can obtain much better results comparing to other similarity based and similarity flooding based algorithms.