Techniques for clustering interaction data as a collection of graphs

TL;DR

This paper presents a novel approach for clustering sequences of graphs derived from interaction data, utilizing model selection, information criteria, and matrix factorization techniques to improve analysis in neuroscience and social networks.

Contribution

It introduces a new method for clustering graphs as a model selection problem, combining information criteria with non-negative matrix factorization and singular value thresholding.

Findings

Effective clustering of graph sequences demonstrated on real data

Improved community detection through graph clustering techniques

Method outperforms existing approaches in simulated experiments

Abstract

A natural approach to analyze interaction data of form "what-connects-to-what-when" is to create a time-series (or rather a sequence) of graphs through temporal discretization (bandwidth selection) and spatial discretization (vertex contraction). Such discretization together with non-negative factorization techniques can be useful for obtaining clustering of graphs. Motivating application of performing clustering of graphs (as opposed to vertex clustering) can be found in neuroscience and in social network analysis, and it can also be used to enhance community detection (i.e., vertex clustering) by way of conditioning on the cluster labels. In this paper, we formulate a problem of clustering of graphs as a model selection problem. Our approach involves information criteria, non-negative matrix factorization and singular value thresholding, and we illustrate our techniques using real and simulated data.