Detection of Potential Citation Clubs in Bibliographic Networks

Abstract

Abstract Citation networks are directed acyclic graphs that are formed by taking research papers as nodes, and two papers are linked if one cites the other giving rise to a directed link. It is well known that there exist groups of authors who collaborate among themselves as well as cite one another’s work. We term such clusters of authors as ‘Potential citation clubs.’ The aim of the paper is to extract the potential citation clubs present in the citation networks and contrast the nature of the clubs that exist in open access journal vs closed access journal. A network called the author-author citation network is constructed from the citation as well as the co-authorship networks. The community discovery algorithms from social network analysis are utilized to detect potential citation clubs in the network. The entire experimentation is carried out using citation and collaboration networks of papers from an open access journal (OJ) and a closed access journal (CJ) from the DBLP database. The citations are limited to papers from the respective journals in order to obtain contrasting trends. The results show that indeed potential citation clubs exist. Further, OJ has clubs with high membership and more closely knit clubs in comparison with those of CJ. More interesting effects may be seen when the experiments are expanded to consider citations to all the journals which is part of our future work.KeywordsCitation networksBibliographic networksCommunity detectionStrongly connected componentsScientometrics