Structure Learning in Bayesian Networks of Moderate Size by Efficient Sampling
This work addresses the challenge of accurate structure learning in Bayesian networks for researchers and practitioners in machine learning and statistics, representing a strong specific gain rather than a broad paradigm shift.
The paper tackles the problem of learning Bayesian network structures from data by developing new algorithms that efficiently sample directed acyclic graphs (DAGs) according to the exact structure posterior, and it shows that these estimators outperform previous state-of-the-art methods in empirical evaluations.
We study the Bayesian model averaging approach to learning Bayesian network structures (DAGs) from data. We develop new algorithms including the first algorithm that is able to efficiently sample DAGs according to the exact structure posterior. The DAG samples can then be used to construct estimators for the posterior of any feature. We theoretically prove good properties of our estimators and empirically show that our estimators considerably outperform the estimators from the previous state-of-the-art methods.