Research into Improved Model for Concept Similarity Computation in Domain Ontology
Yang Mei-ron · Information Sciences · 2014
Currently, Most of concept similarity computation methods are based on ontology structure and IS-A relationship, which discards the semantic information between concepts. To make up the shortage of this method, this paper calculates the Main-Similarity based on ontology structure with the factors of semantic distance between concepts, edge type, concept hierarchy and concept density. Then it uses the relation of attribute in concept similarity computation. Through comparing the similarity of relation names and related concepts between two concepts, it gets influencing factor of attribute relation, with which the Main-similarity be adjusted according to semantics app:ds:semantics. Experimental result shows that this method improves the accuracy of concept similarity computation.