Hierarchical Graph Clustering using Node Pair Sampling

TL;DR

This paper introduces a hierarchical graph clustering algorithm that uses node pair sampling to determine cluster distances, enabling efficient multi-scale graph analysis with proven reducibility and dendrogram output.

Contribution

The paper proposes a novel agglomerative clustering method based on node pair sampling distances, with a proof of reducibility for faster hierarchical clustering.

Findings

Effective on synthetic and real datasets

Produces clear multi-scale graph structures

Speeds up clustering via reducibility proof

Abstract

We present a novel hierarchical graph clustering algorithm inspired by modularity-based clustering techniques. The algorithm is agglomerative and based on a simple distance between clusters induced by the probability of sampling node pairs. We prove that this distance is reducible, which enables the use of the nearest-neighbor chain to speed up the agglomeration. The output of the algorithm is a regular dendrogram, which reveals the multi-scale structure of the graph. The results are illustrated on both synthetic and real datasets.