Structural Analysis of Paper Citation and Co-Authorship Networks using Network Analysis Techniques
Last modified: December 19, 2007
In recent years, attention to structure of existing several complex networks has been increasing. Interesting characteristics of complex networks structure would be applicable in several forms to, for instance, information retrieval and data mining applications. However, in the literature, it has not been sufficiently investigated what the structure of complex networks tells and how the structure of complex networks is applicable. In this paper, we first build a paper citation network and a paper co-authorship network from information on authors and references of a large number of research papers. We then investigate structure of paper citation/co-authorship networks by applying several network analysis techniques, which have been widely used in social network analysis. Our findings include that the paper citation network has less grouped structure than other social networks, and that the tail of the in-degree distribution function follows a power law with $gamma = 2.7$. We also found that that the paper co-authorship network has grouped structure, and exhibits scale-free structure.