基于共现模式的分类号—关键词对应关系研究  

Research on the Corresponding Relationships between Classification Numbers and Keywords Based on Co-occurrence Model

作　　者：钟伟金[1] 

机构地区：[1]广东医学院图书馆,广东湛江520423

出　　处：《情报理论与实践》2013年第5期116-119,共4页Information Studies:Theory & Application

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"的成果;项目编号:10YJC870051

摘　　要：文章以中国生物医学文献数据(CBM)中具有深度标引的分类号及共现的关键词作为统计对象,构建共现矩阵统计三级分类号与关键词的共现频率,在此基础上使用相互包容法计算分类号与关键词共现的紧密度,并以最高紧密度值作为判定关键词所对应的分类号。研究结果表明,该方法有效地将关键词划分到相应的分类号中,准确率达到88.89%,表明这是一种可靠的方法。同时,本文还对四级分类号与关键词的对应关系进行统计分析,结果发现随着分类号的细分,对应关键词的数量减少,显示出一定的层级特性。Taking the classification number with in-depth indexing character and the co-occurrent keywords in the China Bio- logical Medicine （CBM） database as the statistical object, this article constructs a co-occurrence matrix to count up the co-occur- rence frequency of the 3 classes of classification number and the keywords. Then the article uses the mutual tolerance method to cal- culate the tightness of classification number and keyword co-occurrence, and takes the highest tightness value as the corresponding classification number for the judgment of keywords. The research results show that this method can effectively put the keywords into the corresponding classification number, its accuracy reaches 88. 89%, and it is a reliable method. The article also makes a statis- tical analysis of the corresponding relationships between the 4 classes of classification number and the keywords, and the results show that with the subdivision of classification number, the number of the corresponding keywords reduces, revealing a certain lev- el characteristics.

关 键 词：分类法 关键词 对应识别 共现关系 

分 类 号：G353.1[文化科学—情报学]