2022-11-14 Stat Cafe - Dr. Jesus Arroyo Relion

• Time: Monday 11/14 from 11:30 AM to 12:30 PM
• Location: BLOC 503
• Snacks and drinks will be provided
• Gallery

Topic

Community detection in multilayer networks via spectral methods

Abstract

Community detection is the problem of partitioning the vertices of a graph into coherent groups. In this talk, I will present some recent methodologies for community detection in multiple networks. In particular, we study extensions of the (degree-corrected) stochastic block model to a multilayer setting, in which the networks share an underlying community structure but allow for individual graph differences at the local and global levels. We present joint spectral methods that lead to consistent estimation of the community memberships. The methods are effective in handling graph heterogeneity and yield computationally fast and accurate performance in practice. We illustrate the methodology in a time series of flight network data to analyze the dynamics during the covid pandemic.