作　　者： ;

机构地区： 广东医科大学

出　　处： 《情报理论与实践》 2014年第1期131-135,共5页

摘　　要： 文章以学科文献为语料库，构建基于关键词、主题词、副主题词的大型共现词网，分别从同义词、层级关系、相关关系及知识推导四方面实现本体的改造过程。根据相关词标引原理，提炼出“同义相斥、相关相吸”的共现理论，进行同义异形关键词及对应主题词的识别；以《中国分类主题词表》作为工具，实现共现词网主题词的层级关系的标识；以副主题词的组配规则，实现主题词词性的及知识的推导。最终将共现词网改造成集自然语言（关键词）及控制语言（主题词）于一体的领域本体。 Taking the disciplinary literatures as the corpus, this paper constructs a large-sized co-occurrence word net based on keywords, subject headings and sub-subject headings, and implements the ontology transformation process from the aspects of synonyms, hierarchical relations, correlationships and knowledge derivation separately. The paper firstly extracts the co-occurrence theory of ＂synonym mutually repelled, and correlation mutually attracted＂ according to indexing rules of correlated words to identify the abnormal synonymous keywords; then, takes the Chinese Classified Thesaurus as the tool to mark the hierarchical relations of the subject headings in the co-occurrence word net; and lastly uses the coordinate indexing rules of the sub-subject headings to de- rive the part of speech of subject headings and the knowledge. Finally, the paper transforms the co-occurrence word net into the do- main ontology combining the natural language （keywords） and the control language （subject headings） .

关 键 词： 词共现 领域本体 关键词 自动构建

分 类 号： [G254.2]

领　　域： []