ECAI 2004 Conference Paper

[full paper]

Learning Augmented Naive Bayes Classifiers

Josep Roure Alcobé

We propose two general heuristics to transform a batch Hill-climbing search into an incremental one. Our heuristics, when new data arrive, study the search path of the former learning step to determine whether to revise the current structure or not. Heuristics are also able to determine which part of the learned structure should be revised. Then, we apply our heuristics to two well-known Bayesian network structure learning algorithms in order to obtain incremental Augmented Naive Bayes classifiers. We experimentally show that our incremental approach saves a significant amount of computing time while yields classifiers of similar quality than the ones learned with the batch approach.

Keywords: incremental learning, hill-climbing search, Naive Bayes classifiers

Citation: Josep Roure Alcobé: Learning Augmented Naive Bayes Classifiers. In R.López de Mántaras and L.Saitta (eds.): ECAI2004, Proceedings of the 16th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2004, pp.539-543.