扩展主题图本体融合策略与算法  

Strategy and Algorithm for Merging Ontologies of Extend Topic Maps

作　　者：薛咏[1,2] 冯博琴[1] 刘卫涛[1] 

机构地区：[1]西安交通大学电子与信息工程学院,西安710049 [2]西南科技大学信息工程学院,四川绵阳621010

出　　处：《西安交通大学学报》2011年第10期13-18,共6页Journal of Xi'an Jiaotong University

基　　金：国家高技术研究发展计划资助项目(2008AA01Z131)

摘　　要：针对分布式建立与存储的领域本体主题图在融合过程中的语义与结构重复问题以及冗余信息的判断与消除问题,提出了基于语义词典与语料库相结合的主题图融合算法(TMMC),给出了概念相似度计算以及同义关系、整体部分关系等的处理方法.对本体中概念进行基于HowNet语义词典及其他语义词典的多层次相似度计算,定义概念间不同语义关系的融合规则,针对专业领域本体中大量术语词典未收录的问题,提出基于语料库的概念相似度算法,并对计算机教育专业领域扩展主题图进行了融合实验.实验结果表明,TMMC提高了融合的准确率与查全率.Aiming at the distributed design and storage of topic map of domain ontology in abundance latent repetition, and redundancy in meanings and structures during merging process, a topic map merging algorithm（TMMC）hased on common knowledge thesauri combined with corpus, is proposed, where the similarity of Chinese concepts in merging process is evaluated, then the hyponyms, part-of and other relations between concepts, are dealt with, and HowNet and the other thesauri are used to compose a multi-level similarity calculation during merging the concepts in ontology. The merging rules about the different semantic relationships among concepts are de- fined. For very large number of domain special terms are not included in the common knowledge thesauri, therefore a novel algorithm based corpus is presented for calculating similarities between domain special concepts. And the extended topic maps in computer specific education are tested to verify the improved F-measure value of TMMC.

关 键 词：本体融合 相似度计算 语料库 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]