From continuous to discrete variables for Bayesian network classifiers
F. El-Matouat, Olivier Colot, P. Vannoorenberghe, Jacques Labiche · 2002
Using graphical models to represent independent structure in multivariate probability models was studied over a few years. In this framework, Bayesian networks are proposed as an interesting approach for uncertain reasoning. Within the framework of pattern recognition, many methods of classification have been developed based on statistical data analysis. Belief networks were not considered as classifiers until the discovery that Naive Bayes, a very simple kind of Bayesian network, is surprisingly effective. The authors propose the use of belief network classifiers with optimal variables, i.e., networks which have to manage discrete and continuous variables.