Abstract:
|
One of the most relevant features of networks representing real systems is the community structure. Detecting communities is of great importance in understanding, analyzing, organizing networks, as well as in making informed decisions. In recent years, many approaches have been proposed for detecting the community structures in networks. However, few methods have been proposed for testing the statistical significance of detected community structures. In this talk, we describe a statistical framework for modularity based network community detection. Under the proposed framework, a hypothesis testing procedure is developed to determine the significance of an identified community structure. Moreover, the proposed modularity is shown to be consistent under a degree-corrected stochastic block model framework. Several synthetic and real networks are used to demonstrate the effectiveness of our method.
|