A Node Importance Based Label Propagation Approach for Community Detection

Abstract

Community detection provides an important tool to get a deeper insight about the various existing real-world social networks. A large amount of algorithms for detecting community in social networks have been developed in recent years. Most of these algorithms have high computational complex and expensive time-consuming which result that they are not scalable for large-scale networks, and some of them cannot find stable communities. In this paper, we take the importance of node into consideration and give a score to quantize the importance of nodes. Based on this importance quantification, we propose a novel node importance based label propagation algorithm for community detection. We implement our algorithm and other compared algorithms on several benchmark networks. And the experimental results show that our algorithm performs better than others. © Springer-Verlag Berlin Heidelberg 2014.