Not so naive Bayesian classification

Of numerous proposals to improve the accuracy of naive Bayes by weakening its attribute inde- pendence assumption, both LBR and TAN have demonstrated remarkable error performance. How- ever, both techniques obtain this outcome at a con- siderable computational cost. We present a new approach to weakening the attribute independence assumption by averaging all of a constrained class of classifiers. In extensive experiments this tech- nique delivers comparable prediction accuracy to LBR and TAN with substantially improved com- putational efficiency.