一种改进的本体概念语义相似度算法研究  

Research on an Improved Algorithm of Concept Semantic Similarity Based on Ontology

作　　者：刘景方[1] 邹平[1] 张朋柱[2] 齐峰[3] 

机构地区：[1]昆明理工大学管理与经济学院,昆明600093 [2]上海交通大学安泰经济与管理学院,上海200052 [3]浙江财经学院信息学院,杭州310018

出　　处：《武汉理工大学学报》2010年第20期112-117,共6页Journal of Wuhan University of Technology

基　　金：国家自然科学基金(70533030);国家自然科学基金(70471085);浙江省教育厅科技计划项目(Y200803412)

摘　　要：提出了一种改进的基于本体的概念语义相似度算法。文中将传统的基于语义距离的算法和基于信息论的算法进行集成和改进,既克服了信息论方法语义区分不细致的问题,又使得语义距离的计算具有一定的理论严谨性。在分析节点深度、节点密度对于语义相似度计算的影响的基础上,构建节点深度、节点密度2个影响因子以改善概念相似度算法。实验结果表明,改进算法比传统算法具有更好的区分度,并与专家经验吻合较好,得到的语义相似度具有较好的实用价值,有效改善概念语义相似度计算的准确性。This paper presents an improved concept semantic similarity algorithm based on Ontology.The paper has integrated and improved the conventional algorithm based on the semantic distance and the conventional algorithm based on information theory.The algorithm can deal with the problem that the information theory method cannot painstaking distinguish the semantic meaning,and the algorithm make the semantic distance computation to be rigorous.Based on the element analysis of semantic similarity computation,the two impact factors of node distance and node density are built to improve concept semantic similarity algorithm.The experimental results shows that the improved algorithm has better differential capacity than the conventional algorithm,and the result is consistent with experts＇ experience,the semantic similarity has the practical significance,and it can improve efficiently accuracy in the concept semantic similarity computation.

关 键 词：本体 语义相似度 概念 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]