Comparative Study for Inference of Hidden Classes in Stochastic Block Models

TL;DR

This paper compares naive mean field, spectral, and belief propagation methods for inferring hidden classes in stochastic block models, demonstrating belief propagation's superior accuracy and efficiency.

Contribution

It provides a comprehensive comparison of three inference methods, highlighting the advantages of belief propagation over traditional approaches.

Findings

Belief propagation outperforms mean field and spectral methods in accuracy.

Belief propagation is more computationally efficient.

Belief propagation is less prone to overfitting.

Abstract

Inference of hidden classes in stochastic block model is a classical problem with important applications. Most commonly used methods for this problem involve na\"{\i}ve mean field approaches or heuristic spectral methods. Recently, belief propagation was proposed for this problem. In this contribution we perform a comparative study between the three methods on synthetically created networks. We show that belief propagation shows much better performance when compared to na\"{\i}ve mean field and spectral approaches. This applies to accuracy, computational efficiency and the tendency to overfit the data.