ICFC: A Method for Computing Semantic Similarity among Fuzzy Concepts in a Fuzzy Ontology

Lingyu Zhang,Z. M. Ma
DOI: https://doi.org/10.1109/fuzz-ieee.2012.6251166
2012-01-01
Abstract:Semantic similarity calculation is an important step of ontology mapping that is an effective method to solve the problem of ontology heterogeneity. However, current semantic similarity methods can be only adapted to crisp concepts, that is to say, they are not sufficient for handling fuzzy concepts whose instances belong to them with memberships. Therefore, this paper proposes a novel semantic similarity method, ICFC (Information Content of Fuzzy Concept), for fuzzy concepts in a fuzzy ontology. In this method, a semantic similarity between two fuzzy concepts is computed by their information content values. But, it is difficult to obtain information content values for fuzzy concepts. That is because a fuzzy concept is considered as a fuzzy set, and all the elements (i.e., instances) belong to it with memberships. For this purpose, ICFC achieves two tasks: (i) determines whether or not an instance belongs to a fuzzy concept; (ii) provides a method for computing the membership of instance to fuzzy concept, which exploits the degrees that the attribute domains of fuzzy concept include the attribute values of instance. Experimental results with a fuzzy ontology from the real world indicate that ICFC performs encouragingly well.
What problem does this paper attempt to address?