基于维基百科的领域本体自动构建方法研究  

ON WIKIPEDIA-BASED AUTOMATIC DOMAIN ONTOLOGY CONSTRUCTION METHOD

作　　者：吴洁明[1] 刘雁昆 段建勇[1] 

机构地区：[1]北方工业大学计算机学院,北京100144

出　　处：《计算机应用与软件》2016年第7期72-75,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(61103112);国家科技支撑计划项目(2012BA H04F01;2012BAH04F03);北京市科技创新平台基金项目(PXM2013_014212_000011);北京市属高等学校创新团队建设与教师职业发展计划项目(IDHT20130502)

摘　　要：随着互联网的发展,本体被广泛应用于知识工程、信息检索等领域。传统的本体构建方法无法满足日益增长的需求,提出一种基于维基百科的领域本体构建方法。通过提取维基百科的知识体系和知识属性,实现原始领域本体的自动化构建。提出的一种与路径无关的矢量方差方法,通过计算领域中每一个分类和条目的所属度,选择所属度大于阈值的分类和条目作为领域本体的内容,然后挖掘每个信息盒的内容来获取实体的属性。最后通过人工识别的方式验证提取的分类和条目的有效性,以及所构建的本体的准确性。With the development of the Internet,ontology has been widely used in knowledge engineering,information retrieval and other fields. Traditional ontology construction method can't satisfy the growing demand. This paper proposes a Wikipedia-based domain ontology construction method. By extracting the knowledge systems and knowledge attributes of Wikipedia,it realises the automated construction of original domain ontology. In this paper we introduces a path-independent vector variance method,by calculating the belongingness degree of each classification and entry in the field,it chooses those classifications and entries with the belongingness degrees greater than the threshold as the contents of domain ontology,and then mines the contents of each message box to get the attributes of the entity. In end of the paper,through artificial identification way we verify the effectiveness of the extracted classifications and entries and the accuracy of the constructed ontology.

关 键 词：本体 维基百科 有效性 自动构建 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]