A Computation Method of Concept Semantic Similarity in Heterogeneous Ontologies

In allusion to the shortcoming of the existing methods for concept similarity computation that just limited to a certain aspect of concept and can´t fully reflect the degree of concept similarity, this paper presents a calculate method of the concept semantic similarity in heterogeneous ontologies. Joining the concept of bridge-connected, this paper improves the classical computation methods of similarity at present that the distance-based similarity computation method. The improved method takes all the efforts to consider the impact of computation result cased by the semantic distance, concept attributes and the number of causes, and then set different weights to calculate the comprehensive semantic similarity between concepts. Finally compare the computation method with probability statistic method and the human subjective judgment, proving this method is valid.