A framework of an effective fuzzy ontology alignment technique

Abstract

Fuzzy ontology integration is important for handling uncertain information on the Semantic web. However, current ontology integration technologies are not sufficient for fuzzy ontologies. The main contribution of the approach presented here is to propose a novel framework of an effective method for fuzzy ontology alignment. The key concept of the approach is to start from an aligned pair of concepts (called a medoid pair) to determine Potentially Common Parts in different fuzzy ontologies. The distance between the Potentially Common Parts is the sum of the distances between weights of their corresponding concepts. The weight of each concept is estimated by taking into account its attributes and relations to other concepts. Concepts belonging to the common parts are often similar, as they are either sub-concepts/super-concepts of or related concepts to the medoid concepts. Therefore, the distance between Potentially Common Parts is minimized to identify possible matching concepts. New aligned pairs are collected across the potentially common parts by computing similarities between the corresponding concepts belonging to the minimum Potentially Common Parts.