基于共现“互斥互信”原理的同义词识别  

"Mutual exclusion and mutual trust" co-occurrence principle-based identification of synonyms

作　　者：钟伟金[1] 

机构地区：[1]广东医学院图书馆,广东湛江524023

出　　处：《中华医学图书情报杂志》2012年第5期1-4,共4页Chinese Journal of Medical Library and Information Science

基　　金：教育部人文社会科学研究一般项目"共现词汇语义关系挖掘与本体自动构建研究"(10YJC870051);广东高校优秀青年创新人才培育项目"基于词共现的关键词语关系自动识别研究"(wym09089)

摘　　要：在分析关键词、主题词间共现关系的基础上,提出了同义关键词间存在"互斥互信"的理论,并根据这一理论设计了相应的统计模型,从复杂的词间共现关系网中,识别出同义关键词。实验结果表明这种识别方法能有效识别出词形不同的同义词,并具有较高的准确率和可操作性。After the relation between co-occurrence of key words and subject headings was analyzed, the ＂mutual exclusion and mutual trust＂ theory in synonymous key words was proposed, based on which, a corresponding statistical model was designed and synonymous key words were identified from the co-occurrence relation net between complicated words. The results showed that this method could effectively identify the different morphological synonyms with a rather high accuracy and maneuverability.

关 键 词：同义词 自动识别 共现词组 主题词 

分 类 号：G254.2[文化科学—图书馆学] R-058[医药卫生]