Efficient Learning of Selective Bayesian Network Classifiers

In this paper, we empirically evaluate algorithms for learning four types of Bayesian network (BN) classifiers – Naïve-Bayes, tree augmented Naïve-Bayes (TANs), BN augmented Naïve-Bayes (BANs) and general BNs (GBNs), where the GBNs and BANs are learned using two variants of a conditionalindependence based BN-learning algorithm. Based on their performance, we then define a new type of classifier. Experimental results show the resulting classifiers, learned using the proposed learning algorithms, are competitive with (or superior to) the best classifiers, based on both Bayesian networks and other formalisms, and that the computational time for learning and using these classifiers is relatively small. These results argue that BN classifiers deserve more attention in machine learning and data mining communities.