Abstract: The research about concept similarity is very important in knowledge representation and information retrieval.Aiming at the problems of two traditional semantic similarity computation methods，this paper puts forward an integrated method based on ontology.This method integrates the graph structure of ontology with several kinds semantic influence factors and utilizes the concept semantic information fully to compute the semantic similarity between concepts.The experiment shows that the computation result is reasonable and this method is effective.

Key words: ontology, semantic similarity, semantic distance, information content, common disjunction ancestor

摘要： 概念的语义相似度研究，是知识表示以及信息检索领域中的一个重要内容。通过分析两种传统的语义相似度计算方法，对它们存在的问题进行改进，提出了一种综合的基于本体的概念语义相似度计算方法。该方法结合本体的DAG网状结构特征和语义距离计算中的多种语义影响因素，充分利用本体中概念的语义来计算概念间的语义相似度。实验结果比较合理，验证了该方法的有效性。

关键词: 本体, 语义相似度, 语义距离, 信息量, 共同分离祖先