基于复杂网络的本体结构分析  

Complex Network-based Ontology Structure Analysis

作　　者：张大陆[1] 王志晓[1,2] 刘雯[1] 杨哲[1] 

机构地区：[1]同济大学计算机科学与工程系,上海201804 [2]中国矿业大学计算机学院,江苏徐州221116

出　　处：《同济大学学报（自然科学版）》2009年第2期258-261,共4页Journal of Tongji University:Natural Science

基　　金：国家自然科学基金资助项目(90202010)

摘　　要：将本体的结构特点与网络类比,本体中的概念与网络中的点对应,本体中的关系与网络中的边对应,借助现有的针对复杂网络的分析方法和性能指标,对本体的度分布、平均最短路径、聚集系数进行统计分析,考察本体结构特性.选取基因本体(GO)作为研究样本,分析结果表明,GO没有很高的聚类特性,具有小世界特性,不具有无尺度性质.该研究有助于评估本体中的概念,设置其权重,提高语义匹配的效率.Concepts and relations between concepts in ontology can be considered as the vertex and edges in network due to their similarity in structure. As a result, the analyzing method and metrics for complex network may also be used to define the complexity of ontology. In the study the degree distribution, clustering coefficient and average distance of GO ontology are measured to analyze its complexity. The result shows that the GO ontology has small-world property but does not have scale free property. There is no clustering characteristic in GO ontology. The study can be used to evaluate importance of ontology concepts and to improve the efficiency of semantic matching.

关 键 词：复杂网络 本体 聚集系数 小世界 无尺度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]