Key community analysis in scientific collaboration network

Abstract

Till date, the community evaluation in the field of scientific collaboration network, have been taken place on an individuals' impact. Very less findings have been done on the community detection and community evaluation considering the acquaintances of co-authors. This paper comprises of network in which the database of author's publications is used. The authors are the nodes and they are connected because they co-authored a scientific paper. We used these databases to answer questions about the key community in the collaboration network and led the findings of network centralization, network density and average cluster co-efficient of the scientific community detected using the concepts of node similarity, node degree and node reachability. We detected top 20 communities and computed network centralization, density and average clustering co-efficient. After analysis, we realized that, as the total number of nodes increases within a cluster, the lesser or equivalent are the values for network centralization, density and average cluster coefficient.