Hierarchical Community Evolution Mining from Dynamic Networks

Research on community evolution contributes to understanding the nature of network evolution. Previous community evolution studies have two defects: (1) the algorithms do not have sufficient stability or cannot handle the radical structure change of communities, and (2) they cannot reveal the evolutionary regularities with multiple levels. To solve these problems, this paper proposes a new method for mining the evolution of communities from dynamic networks. Experiments demonstrate that compared with traditional methods, our work significantly improves the algorithm performances.