Seol mar théacs é seo: Discretizing environmental data for learning Bayesian-networkclassifiers