A Bayesian Network Classifier that Combines a Finite Mixture Model and a NaIve Bayes Model

In this paper we present a new Bayesian net­ work model for classification that combines the naive Bayes (NB} classifier and the fi­ nite mixture (FM} classifier. The resulting classifier aims at relaxing the strong assump­ tions on which the two component models are based, in an attempt to improve on their classification performance, both in terms of accuracy and in terms of calibration of the estimated probabilities. The proposed clas­ sifier is obtained by superimposing a finite mixture model on the set of feature variables of a naive Bayes modeL We present exper­ imental results that compare the predictive performance on real datasets of the new clas­ sifier with the predictive performance of the NB classifier and the FM classifier.