The Grow-Shrink strategy for learning Markov network structures constrained by context-specific independences

TL;DR

This paper introduces CSGS, a new algorithm for learning Markov network structures that incorporate context-specific independences, reducing computational complexity while maintaining accuracy.

Contribution

It proposes CSGS, a Grow-Shrink based algorithm that improves efficiency in learning Markov network structures with context-specific independences.

Findings

CSGS achieves comparable accuracy to CSPC.

CSGS reduces computational complexity significantly.

Empirical results validate the effectiveness of CSGS.

Abstract

Markov networks are models for compactly representing complex probability distributions. They are composed by a structure and a set of numerical weights. The structure qualitatively describes independences in the distribution, which can be exploited to factorize the distribution into a set of compact functions. A key application for learning structures from data is to automatically discover knowledge. In practice, structure learning algorithms focused on "knowledge discovery" present a limitation: they use a coarse-grained representation of the structure. As a result, this representation cannot describe context-specific independences. Very recently, an algorithm called CSPC was designed to overcome this limitation, but it has a high computational complexity. This work tries to mitigate this downside presenting CSGS, an algorithm that uses the Grow-Shrink strategy for reducing unnecessary computations. On an empirical evaluation, the structures learned by CSGS achieve competitive accuracies and lower computational complexity with respect to those obtained by CSPC.