%0 Journal Article
%T 语义网虚拟本体构建<br>Building virtual ontologies in semantic web
%A 张祥
%A 李星
%A 温韵清
%A 沈凯
%A 郝敬坤
%J 东南大学学报（自然科学版）
%D 2015
%R 10.3969/j.issn.1001-0505.2015.04.007
%X 针对缺乏良好组织的本体难以被重用的问题,提出了对本体中术语进行语义重组的方法.该方法首先从本体文档中抽取类之间的关系,建立类关系图,进而对不同类型的关系赋以权重.然后将多本体中语义高度相关的术语进行组合,对语义相关度低的术语进行分割.最后分别通过基于划分、基于层次化的聚类算法将类关系图重组为一组虚拟本体,实现了对本体的有效重用.以语义搜索引擎Falcons索引的本体作为数据集,对生成的虚拟本体进行了时间性能和聚类合理性评估,并对比了2种聚类方法在虚拟本体生成中的作用.实验证明该方法生成的虚拟本体具有较好的语义内聚性,达到了对本体中术语进行语义重组的目的.<br>As the disorganized ontologies are prone to bringing barrier to ontology reuse, a novel approach to semantically re-organizing terms in multiple ontologies is proposed. Relations between classes are extracted from ontological documents to build a class relation graph, and different types of relations are weighted. Semantically related terms from multiple ontologies are combined into a virtual ontology, and unrelated terms are seperated. Virtual ontologies are constructed by performing partition-based and hierarchical clustering algorithms on class relation graph, respectively, and the virtual ontologies can be effectively reused. Ontologies in a semantic search engine, Falcons, are used as dataset for evaluation. Time performance and the rationality of clustering are evaluated, and the two clustering algorithms are compared by their effectiveness in the construction of virtual ontologies. Experiments show that virtual ontologies produced by the proposed approach are semantically cohesive, and the approach can semantically re-organize terms in ontologies
%K 本体
%K 虚拟本体
%K 语义网
%K 聚类&lt
%K br&gt
%K ontology
%K virtual ontology
%K semantic web
%K clustering
%U http://journal.seu.edu.cn/oa/darticle.aspx?type=view&id=201504007