CCLPA: A clustering coefficient based label propagation algorithm for unfolding communities in complex networks

Identifying interconnected groups in complex networks such as social networks, biological networks, and communication networks is an ever important task in data analysis. These interconnected groups are termed as communities in social network analysis and plays an important role in understanding the structural and behavioral properties of complex networks. In this paper, we propose a novel label propagation algorithm, called CCLPA (A Clustering Coefficient based Label Propagation Algorithm) to address the randomness issue of label propagation algorithm. Our algorithm defines the function, clustering coefficient, to measure the neighborhood connectivity between nodes quantitatively without any contact with the user. Based on the clustering coefficient, we present a new label propagation algorithm with explicit node update sequence to uncover communities in complex networks. Experiments on real-world network datasets demonstrate that it overcomes the random initial label selection and random label update order of underlying label propagation algorithm. Our algorithm identifies stable communities and becomes more robust and efficient. Wide experiments show the better-quality and effectiveness of the proposed algorithm.