作者文献耦合分析与作者关键词耦合分析比较研究：Scientometrics实证分析

A Comparative Study on Author Bibliographic-coupling Analysis and Author Keyword-coupling Analysis Based on Scientometrics

本文以Scientometrics 1991—2010年间刊载的2,045篇论文作为数据样本,分两个时间段研究这20年间科学计量学的知识结构与演进状况,并以此来研究作者文献耦合分析法(ABCA)与作者关键词耦合分析法(AKCA)在揭示学科领域知识结构方面的异同。研究发现,2001—2010年间的科学计量学的知识结构要比1991—2000年间的更加清晰明朗,其研究主题之间更为亲密,相互作用力明显要强。作者排名相关分析、研究主题探测、余弦相似度计算、研究主题变迁等均显示,AKCA与ABCA存在高度相关性;通过因子分析的模型拟合、研究主题的探测与变迁分析,又显示二者略有不同,ABCA可以探寻到比AKCA更多的研究主题,AKCA比ABCA能显示更多的信号来反映学科的技术突破以及研究前沿的进展。因此ABCA与AKCA不可互相替代,二者结合起来是探寻学科知识结构及其发展的理想研究方法。

Taking 2,045 papers published in Scientometrics from 1991 to 2010 as a sample and by dividing them into 2 groups according to different time periods, we study the knowledge structure and evolution of Scientometrics during these 20 years, and then the two methods for disciplinary structure analysis： author bibliographic-coupling analysis（ABCA） and author keyword-coupling analysis（AKCA）. We find that the knowledge structure of Scientometrics between 2001 and 2010 is more clear than that between 1991 and 2000, the relationship among research topics of the former is closer and the interaction force of the former is stronger. Related analysis of authorranking, exploration of reserach topics, calculation of cosine similarity and changes of research topics all show that there are both high correlation and slight difference between ABCA and AKCA with the former exploring more research topics and the latter displaying more signals to reflect technalogical breakthroughs and the development of research front in the disciplines. A combination of the two methods is proposed in exploring＇ disciplinary structure and devel- opment 4 figs. 6 tabs. 25 refs.