Locally Minimax Optimal Predictive Modeling with Bayesian Networks
Abstract
We propose an information-theoretic approach for predictive modeling with Bayesian networks. Our approach is based on the minimax optimal Normalized Maximum Likelihood (NML) distribution, motivated by the MDL principle. In particular, we present a parameter learning method which, together with a previously introduced NML-based model selection criterion, provides a way to construct highly predictive Bayesian network models from data. The method is parameter-free and robust, unlike the currently popular Bayesian marginal likelihood approach which has been shown to be sensitive to the choice of prior hyperparameters. Empirical tests show that the proposed method compares favorably with the Bayesian approach in predictive tasks.
Cite
Text
Silander et al. "Locally Minimax Optimal Predictive Modeling with Bayesian Networks." Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics, 2009.Markdown
[Silander et al. "Locally Minimax Optimal Predictive Modeling with Bayesian Networks." Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics, 2009.](https://mlanthology.org/aistats/2009/silander2009aistats-locally/)BibTeX
@inproceedings{silander2009aistats-locally,
title = {{Locally Minimax Optimal Predictive Modeling with Bayesian Networks}},
author = {Silander, Tomi and Roos, Teemu and Myllymäki, Petri},
booktitle = {Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics},
year = {2009},
pages = {504-511},
volume = {5},
url = {https://mlanthology.org/aistats/2009/silander2009aistats-locally/}
}