A method to improve the time of computing for detecting community structure in social network graph

Abstract

Identifying communities has always been a fundamental task in analysis of complex networks. Currently used algorithms that identify the community structures in large-scale real-world networks require a priori information such as the number and sizes of communities or are computationally expensive. Amongst them, the label propagation algorithm (LPA) brings great scaslability together with high accuracy but which is not accurate enough because of its randomness. In this paper, we study the equivalence properties of nodes on social network graphs according to the labeling criteria to shorten social network graphs and develop label propagation algorithms on shortened graphs to discover effective social networking communities without requiring optimization of the objective function as well as advanced information about communities. Test results on sample data sets show that the proposed algorithm execution time is significantly reduced compared to the published algorithms. The proposed algorithm takes an almost linear time and improves the overall quality of the identified community in complex networks with a clear community structure.