基于Stanford TMT的国际智库研究内容挖掘分析

摘要

Think tanks research is a hot point in many disciplines. Through an analysis of 727 research articles on this domain from Web of Science database by using Stanford Topic Modeling Toolbox (TMT) for content mining, it is found that the international think tank research are clustered into two groups: construction and development research of think tanks, and applications of think tanks with 10 subtopics. The Stanford TMT which is used for topic model analysis can detect latent semantic topics of articles effectively and enhance the bibliometric methods such as co-words analysis.%智库研究是当前各学科领域关注的共同热点。文章通过搜集WOS数据库中与智库相关的727篇研究论文，利用Stanford TMT主题模型工具包对研究论文的文本进行了主题挖掘，分析结果表明，当前国际智库研究主要包括智库建设与发展研究、特定领域和问题解决中的智库研究等两大类10个主题；通过本研究的主题挖掘结果，也证明利用Stanford TMT主题模型工具包能够有效挖掘出文献隐含的语义主题，主题模型分析能够对共词分析等文献计量方法进行良好的补充。