Analyzing the Impact of the Discretization Method When Comparing Bayesian Classifiers

Most, of the 1110) 110(15 designed within 1,the framework of Bayesian net networks (BNs) assume that the involved variables ale Of discrete nature but this assumption rarely holds in real problem. The Bayesian classifier AODE (Aggregating One-Dependence Estimators) e g can only classifier directly with discrete variables The HAODE Hybrid AODE) classifier is proposed as an appealing alternative to 11 ODE which is less affected by the discretization process In this paper, study if this behavior holds when applying different discretization methods More importantly, (ye include other Bayesian classifiers in the comparison to find out. to., what extent, the type of discretization affects then results in terms of accuracy and bias-variance discretization If the type of discretization applied is not decisive, then future expel intents can be A: times faster, being the number of discretization methods considered.