改进的基于本体的语义相似度计算  

Improved ontology-based semantic similarity computation algorithm

作　　者：张沪寅[1] 温春艳[1] 刘道波[1] 叶刚[1] 

机构地区：[1]武汉大学计算机学院,湖北武汉430072

出　　处：《计算机工程与设计》2015年第8期2206-2210,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61272454);国家教育部博士点基金项目(20130141110022)

摘　　要：针对已有的基于本体的相似度算法存在各类多继承和准确度不高的问题,提出一种改进的综合加权的相似度计算算法PRSSC。在共享路径重合度的基础上进行改进,综合加权概念节点密度、深度、最低共同祖先节点深度以及概念的属性。实验结果表明,与现有相似度算法比较,PRSSC算法可在解决各种多继承问题的基础上,提高相似度计算结果准确度。Analyzing the disadvantages of the current ontology-based similarity calculation algorithm, such as the problem of multiple inheritance and low accuracy of the calculation results, an improved integrated weighted algorithm PRSSC was pro- posed. Algorithm was improved on the basis of the contact ratio with the shared path. Then weighting the node density, depth, depth of the lowest common ancestor nodes （LCA）, and the property of the concepts were integrated. Compared with the exis- ting similarity algorithms, the results show that the PRSSC algorithm can increase the accuracy of similarity results and solve va rious problems of the multiple inheritances.

关 键 词：本体 语义相似度 共享路径 概念深度 多继承 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]