GOBNILP: Learning Bayesian network structure with integer programming
Proceedings of the 10th International Conference on Probabilistic Graphical Models, PMLR 138:605-608, 2020.
The GOBNILP system for learning Bayesian networks is presented. Both the Python and C implementations are discussed. The usefulness of learning multiple BNs is highlighted. Current work on ‘pricing in’ new integer programming variables is presented.