Learning Sparse Graphs with a Core-periphery Structure

TL;DR

This paper introduces a generative model to learn sparse graphs with a core-periphery structure, jointly inferring the graph and core scores from node attributes, demonstrating effectiveness on real-world data.

Contribution

It proposes a novel generative model that jointly infers sparse core-periphery structured graphs and core scores from node attributes, advancing graph learning methods.

Findings

Successfully learns core-periphery structure from node attributes alone

Core score estimates align well with existing methods that use graph data

Demonstrates effectiveness on diverse real-world datasets

Abstract

In this paper, we focus on learning sparse graphs with a core-periphery structure. We propose a generative model for data associated with core-periphery structured networks to model the dependence of node attributes on core scores of the nodes of a graph through a latent graph structure. Using the proposed model, we jointly infer a sparse graph and nodal core scores that induce dense (sparse) connections in core (respectively, peripheral) parts of the network. Numerical experiments on a variety of real-world data indicate that the proposed method learns a core-periphery structured graph from node attributes alone, while simultaneously learning core score assignments that agree well with existing works that estimate core scores using graph as input and ignoring commonly available node attributes.