作者合著网络中研究兴趣相似性实证研究

Empirical Research on Similarity of Research Interests in Co-authorship Network

[目的/意义]从作者微观个体研究兴趣角度出发,通过对作者合著网络中作者关联关键词集的研究,定量地验证研究兴趣相似是作者合作的一个动机。[方法/过程]收集WOS中检索领域相关文献题录信息,构建作者合著网络,并利用Louvain算法划分社区,实现了Jaccard系数及余弦相似性系数的计算指标,统计与对比分析整体网络及社区内部作者研究兴趣的相似性。[结果/结论]在网络整体层次,作者合著网络中作者的研究兴趣相似性较高,但也存在一定比例的差异性即互补性;在科研社区内部,合著作者平均研究兴趣相似性及互补性均高于网络整体层次,科研社区的形成受到作者研究兴趣的影响。两个层次的兴趣相似性反映了研究兴趣相似是作者合作的一个重要动机。

[ Purpose/significance ] Aiming at quantitatively verifying the assumption that the similarity of research in- terests is a motivation of the co-authoring, this paper uses the keyword set of those authors of the co-authorship network from the micro perspective of individual. [ Method/process] The research collects bibliographic record of document about information retrieval, constructs co-authorship network, divides the communities with Louvain algorithm, implements Jac- card coefficient and Cosine coefficient computation indicators, and carries out a statistical analysis and comparison on the similarity of research interests of authors in the overall network and within the community. [ Result/conclusion ] From the perspective of overall network, the similarity of research interest is higher in co-authorship network, but there is a rate of complementarity. Both the average similarity and complementarity of co-author within community are higher than overall network, and the formation of scientific community may be influenced by the research interest of author. That the similari- ty of research interest is an important motivation of the co-authoring can be reflected by the similarity of researcher＇ s inter- est in these two levels of network.