Abstract: To enhance information retrieval recall and precision, this paper proposes an improved method of calculating ontology semantic similarity. This method adds three kinds of weighting factors in the calculation of semantic distance, they are depth, density and type. In addition, it also considers the influence of semantic contact ratio and attributes of the concept. Experimental results show that the method is more accurate and efficient than the traditional calculation method.

Key words: semantic similarity, ontology, information retrieval, semantic distance

摘要： 为提高信息检索的查全率与查准率，提出一种改进的本体语义相似度计算方法。该方法在语义距离的计算中加入了深度、密度、类型三种权重因子，并且综合考虑语义重合度、概念的属性对相似度的影响。通过实验分析，该方法比传统计算方法更加准确、有效。

关键词: 语义相似度, 本体, 信息检索, 语义距离