The most common way of automatically learning Bayesian networks from data is the combination of a scoring metric, the evaluation of the fitness of any given candidate network to the data base, and a search procedure to ...