Community Detection Algorithm Based on the Scale Free Property of Networks

Many complex networks in the real world demonstrate similar patterns, including the scale-free property and strong community structure. In this paper, we present a novel parameter-free community detection algorithm based on the scale-free property of networks, named ScaleFreeCDA. The basic idea behind it is two mechanisms, i.e., node growth and preferential attachment. Community centers are firstly determined by node degree and similarity, and then community structures are obtained accordingly. The experiments indicate that our algorithm can find high-quality communities, and can determine appropriate number of communities on most artificial and real networks.