Mining social networks on the mexican computer science community

Abstract

Scientific communities around the world are increasingly paying more attention to collaborative networks to ensure they remain competitive, the Computer Science (CS) community is not an exception. Discovering collaboration opportunities is a challenging problem in social networks. Traditional social network analysis allows us to observe which authors are already collaborating, how often they are related to each other, and how many intermediaries exist between two authors. In order to discover the potential collaboration among Mexican CS scholars we built a social network, containing data from 1960 to 2008. We propose to use a clustering algorithm and social network analysis to identify scholars that would be advisable to collaborate. The idea is to identify clusters consisting of authors who are completely disconnected but with opportunities of collaborating given their common research areas. After having clustered the initial social network we built, we analyze the collaboration networks of each cluster to discover new collaboration opportunities based on the conferences where the authors have published. Our analysis was made based on the large-scale DBLP bibliography and the census of Mexican scholars made by REMIDEC. © 2009 Springer-Verlag Berlin Heidelberg.