Faster Clustering via Non-Backtracking Random Walks

TL;DR

This paper introduces VEC-NBT, an improved graph clustering method that uses non-backtracking random walks to achieve better accuracy and efficiency, especially on sparse graphs, by modifying the original VEC algorithm.

Contribution

It proposes a novel variation of VEC called VEC-NBT that employs non-backtracking random walks, enhancing clustering performance on sparse graphs.

Findings

VEC-NBT achieves higher accuracy than VEC on sparse graphs.

Shorter random walks are sufficient for VEC-NBT to perform well.

VEC-NBT significantly outperforms VEC in sparse graph clustering.

Abstract

This paper presents VEC-NBT, a variation on the unsupervised graph clustering technique VEC, which improves upon the performance of the original algorithm significantly for sparse graphs. VEC employs a novel application of the state-of-the-art word2vec model to embed a graph in Euclidean space via random walks on the nodes of the graph. In VEC-NBT, we modify the original algorithm to use a non-backtracking random walk instead of the normal backtracking random walk used in VEC. We introduce a modification to a non-backtracking random walk, which we call a begrudgingly-backtracking random walk, and show empirically that using this model of random walks for VEC-NBT requires shorter walks on the graph to obtain results with comparable or greater accuracy than VEC, especially for sparser graphs.