A hybrid strategy for community detection of complex networks

Many complex systems represent modular feature in their structure and evolution mechanisms. This paper reviews state-of-the-art in community detection algorithms and quality measures. On the basis of that, a hybrid strategy for community detection is proposed. First, the lead vertices of the network is folded and the central vertices are emerged through the application of k-core. Then the central vertices are divided into communities based on modularity optimization. The experimental results suggest that the hybrid strategy can produce detection results with larger value of modularity, and the partitioning is well performed in practices.