基于期刊论文的学科间交叉主题识别研究

Research on Interdisciplinary Cross-topic Recognition Based on Journal Articles

本文以知网数据库中管理工程和情报学共12本CSSCI期刊5年的论文作为数据,通过AT主题模型、相似度计算识别学科间交叉主题并对其相似程度进行测度。实验表明,AT主题模型可以挖掘两个学科间交叉研究主题,并提供一定的语义描述。通过对比关键词共现的方法,本文的方法可以提供交叉研究主题更多的语义特征,并便于判断高相似度的研究主题。

A total of 12 CSSCI journals in CNKI about Management Engineering and Information Science are used as the data foundation.Using AT topic model and similarity calculation,this paper identifies the interdisciplinary cross-topics and measures their similarity.Experiments show that AT topic model can mine the cross-topics between two disciplines and provide a certain semantic description.By comparing the methods of keyword co-occurrence,the method of this paper can provide more semantic features for cross-topics and measure high similarity research topicseasily.