Multi Criteria-Based Community Detection and Visualization in Large-scale Networks Using Label Propagation Algorithm

Abstract

Networks in the Big Data era are characterized by complex structures due to their heterogeneity, largeness and dynamics. As result, many issues regarding scalability have emerged. Among them, the community detection problem takes an important part. In the case of large-scale graphs, this problem presents a real issue because of its high complexity and therefore slowness. In this paper, we introduce a new approach based on Label Propagation Algorithm (LPA) to consider the community detection problem in a distributed and scalable way. It can be used for both single and multi-label networks. The experimentation is realized using the Spark GraphX framework. The results show its benefits.