Madaan, GauravGauravMadaanJolad, ShivakumarShivakumarJolad2025-08-302025-08-302014-01-01[9781479956654]10.1109/BigData.2014.70043462-s2.0-84921751665https://d8.irins.org/handle/IITG2025/21555We study the structure and evolution of scientific collaboration network by using collaboration network constructed from DBLP Computer Science Bibliographic database [1], from year 1936 to 2013 using social network analysis techniques. We have found many interesting features such as collaboration between scientists is increasing with time and few numbers of scholars publish a large number of papers while most of the authors publish a small number of papers, which is consistent with Lotka's law on frequency of publications [2]. The degrees of the vertices in the collaboration graph follow a 'Power law' pattern i.e., the number of vertices of degree x is proportional to a negative power of x. The clustering coefficient of collaboration graph comes out to be very high which means that there are more chances for two authors to co-author a paper if they have a common collaborator. We also found that the collaboration graph follows various real graph properties like WPL (Weight power law), DPL (Densification power law) etc. We try to apply the Lorenz curve and Gini coefficient on the collaboration graph to study the variation in concentration of collaboration between researchers with time.falsebibiliometrics | collaboration networks | network structure | power laws | research collaboration | scale free networksConference Paper7-13201497004346cpConference Proceeding7