摘要
本文提出用代表性论文来表示某～学科主题高频主题词共现聚类分析类团内容的设想，并加以验证。代表性论文就是对共现聚类形成贡献最大的前2～3篇论文。以1996年发表的一篇共词聚类分析人工评判结果为基准，将本次研究确认的代表性论文与之对比后，发现代表性论文不仅能完全表达了高频词聚类分析形成类目的内容，而且可以提供更加丰富的主题词间语义关系等内容。这种方式为进一步实现共现聚类分析流程的规范化和自动化提供了有力支持。
This paper presented and testified a hypothesis that a representative papers could be used as a label of a coword clustering results in scientific field. These representative papers are those top 2 or 3 papers mostly contributed to the cluster forming. This paper took the results from a co-word clustering analysis published in 1996 as the baseline, compared the content of the representative papers by present study with the manual analysis content made in 1996, the result showed that sentences and terms from the representative papers cover every concept described in manual analysis, and even provide semantic relations among these high frequency subject headings. This approach provided a firm support to the formalization and automation for the co-occurrence clustering analysis procedure.
出处
《情报学报》
CSSCI
北大核心
2015年第12期1270-1277,共8页
Journal of the China Society for Scientific and Technical Information
关键词
共词分析
聚类分析
类团内容
代表性论文
co-word analysis, clustering analysis, cluster label, representative papers
