Template-Based Graph Clustering

TL;DR

This paper introduces a graph clustering method that uses template matching and prior information to improve clustering accuracy, especially in difficult cases, by embedding graphs into a lower-dimensional space.

Contribution

It presents a novel template-based graph clustering approach that incorporates prior structural information and relaxes the problem to orthonormal matrix optimization.

Findings

Outperforms classical clustering methods in challenging scenarios

Utilizes template matching to guide clustering process

Employs a relaxed optimization framework for embedding

Abstract

We propose a novel graph clustering method guided by additional information on the underlying structure of the clusters (or communities). The problem is formulated as the matching of a graph to a template with smaller dimension, hence matching $n$ vertices of the observed graph (to be clustered) to the $k$ vertices of a template graph, using its edges as support information, and relaxed on the set of orthonormal matrices in order to find a $k$ dimensional embedding. With relevant priors that encode the density of the clusters and their relationships, our method outperforms classical methods, especially for challenging cases.