Title
Community Detection and Extraction in Networks
Speaker
Yunpeng Zhao, University of Michigan
Abstract
Community detection is a fundamental problem in network analysis, with applications in many diverse areas, including computer science, social sciences and biology. This talk contains both theoretical and methodological works in this area. The stochastic block model is a common tool for model-based community detection, but is limited by its assumption that all nodes within a community are stochastically equivalent, and provides a poor fit to networks with hubs or highly varying node degrees within communities, which are common in practice. The degree-corrected block model was proposed to address this shortcoming, which allows variation in node degrees within a community. The first part of the talk will present general theory for checking consistency of community detection under the degree-corrected block model. The second part is about a new community extraction framework, which allows both tight communities and weakly connected background nodes. The proposed extraction criterion performs well on simulated and real networks. We also establish its asymptotic consistency for the case of block model.
