Learning Bayesian Networks from Incomplete Data with Stochastic Search Algorithms

TL;DR

This paper introduces stochastic search algorithms, including a new method with an adaptive mutation operator, for learning Bayesian networks from incomplete data, effectively exploring complex solution landscapes beyond local maxima.

Contribution

The paper presents a novel stochastic search approach with an adaptive mutation operator for Bayesian network learning from incomplete data, addressing limitations of deterministic methods.

Findings

Stochastic algorithms produce accurate Bayesian network structures.

The new adaptive mutation operator improves search effectiveness.

Methods outperform traditional deterministic approaches in exploring solution space.

Abstract

This paper describes stochastic search approaches, including a new stochastic algorithm and an adaptive mutation operator, for learning Bayesian networks from incomplete data. This problem is characterized by a huge solution space with a highly multimodal landscape. State-of-the-art approaches all involve using deterministic approaches such as the expectation-maximization algorithm. These approaches are guaranteed to find local maxima, but do not explore the landscape for other modes. Our approach evolves structure and the missing data. We compare our stochastic algorithms and show they all produce accurate results.