Community Detection in Scientific Co-Authorship Networks using Neo4j

Abstract

Community structure in scientific collaboration network has become an important research area. Coauthor of a paper can be thought of as a collaborative document between more than one authors. Community detection in co-authorship network reveals characteristic patterns of scientific collaboration in computer science research and help to understand the identity-organization of the author community. Louvain algorithm is a simple, easy to implement and efficient to recognize community in huge networks. In this paper, it is used to examine the structure of community in Computer University’s coauthor network in Myanmar. Neo4j is also used to visualize the coauthorship network analysis results. Modularity is used to measure the quality of the cluster structure found by community discovery algorithms. In experiment, Louvain algorithm gives more effective qualitative community structures than other algorithms in co-authorship network.