pith. sign in

arxiv: 1310.1297 · v5 · pith:IHDJCZMLnew · submitted 2013-10-04 · 📊 stat.ML · math.OC· stat.CO

Spectral Clustering for Divide-and-Conquer Graph Matching

classification 📊 stat.ML math.OCstat.CO
keywords graphmatchinggraphslargeveryalgorithmapproachdivide-and-conquer
0
0 comments X
read the original abstract

We present a parallelized bijective graph matching algorithm that leverages seeds and is designed to match very large graphs. Our algorithm combines spectral graph embedding with existing state-of-the-art seeded graph matching procedures. We justify our approach by proving that modestly correlated, large stochastic block model random graphs are correctly matched utilizing very few seeds through our divide-and-conquer procedure. We also demonstrate the effectiveness of our approach in matching very large graphs in simulated and real data examples, showing up to a factor of 8 improvement in runtime with minimal sacrifice in accuracy.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.