CMStatistics 2020: Start Registration
View Submission - CMStatistics
Title: Salient structure identification in complex networks by spectral periphery filtering Authors:  Ruizhong Miao - University of Virginia (United States)
Tianxi Li - University of Virginia (United States) [presenting]
Abstract: Complex networks have been intensively studied in the past fifteen years. In practice, the salient network structure of interest, instead of being directly observed, is often hidden in a larger network in which most structures are not informative. The noise and bias introduced by this overwhelming yet non-informative data can obscure the salient structure and limit the effectiveness of many network analysis methods. Traditionally, researchers treat this scenario as a core-periphery structure, and algorithms are designed to extract the core. Unfortunately, most of these methods rely on restrictive assumptions on both the core and the periphery components that seriously undermine their usefulness. We propose a random network model for the non-informative structure of networks without imposing a specific form for the core. Specifically, we assume that the non-informative nodes are connected to other nodes in a purely random pattern, while the core structure can take any informative pattern. Moreover, we propose an algorithm of core extraction. The algorithm is computationally efficient and comes with a theoretical guarantee of accuracy. We evaluated the proposed model in extensive simulation studies and also use it to extract core structures in a few real-world networks for downstream analysis.