作　　者： ;

机构地区： 广东医科大学

出　　处： 《情报学报》 2009年第2期 214-219,共6页

摘　　要： 共词聚类分析法在词对共现频率的基础上,采用聚类的算法把词（通常是主题词）聚集成类,从而揭露隐含在文献群中知识。词对共现频率结果的准确性、科学性是共词聚类分析成功的关键。本文对共词聚类分法中的词对共现频率进行深入的分析,认为在文献的标引中存在主要主题词与次要主题词的差别,在词对共现频率计算时应对主要主题词进行加权计算,从而突出主要主题词在聚类过程中的主导地位。通过实例的对比分析,说明这种加权对共词聚类分析法的改进是有效的,也是有必要的。 Co-word cluster analysis is a text mining technology,with which phrases（descriptors）are divided into phrasegroup based on clustering algorithms.From the phrase-group,some implicit information of literatures can be exposed.This paper analyzes the characteristics of co-words Frequency about co-word cluster analysis.It is necessary to make major Descriptor weighted calculation for co-word frequency and to highlight the dominant position of major Descriptor in the co-word cluster process because of the differences between major Descriptor and Secondary Descriptor in literature indexing.Comparative analysis of examples demonstrates that the weighted co-word Cluster analysis based on the Major Descriptor is not only effective,but also necessary.

关 键 词： 共词分析 词频研究 主题词 实例研究

分 类 号： [G254.2 F222.1]

领　　域： [] []