The Research on Concept Semantic Similarity Computing Based on Semantic Tree

The concept semantic similarity computation has vast application in the question answering system field, semantic similarity is becoming the basis of semantic expansion in the question answering system, and influences the recall rate and precision rate of the question answering system immensely. After we analyze the traditional method of semantic similarity computation, we consider comprehensively the important influential factors: depth, density, strength of the edge and concept semantic overlap degree, concept level difference. The paper presents a semantic tree-based concept semantic similarity computation method, and verifies the algorithm´s validity.