Community detection in hypernetwork via Density-Ordered Tree partition

Qing Cheng, Zhong Liu, Jincai Huang, Guangquan Cheng

Abstract: Hypernetwork, as a useful representation of natural and social systems has received increasing interests from researchers. Community is crucial to understand the structural and functional properties of the hypernetworks. Here, we propose a new method to uncover the communities of hypernetworks. We construct a Density-Ordered Tree (DOT) to represent original data by combining density and distance, and the community detection in hypernetwork is converted to a DOT partition problem. Then, an anomaly detection strategy using box-plot rule is applied to partition DOT and judge whether there is a significant community structure in the hypernetwork. Moreover, visual inspection as a complementary approach of box-plot rule can effectively improve the effectiveness of community detection. Finally, the method is compared with existing methods in both synthetic and real-world networks.