Abstract: We consider estimation of Bayesian network structures given a finite number of examples when both discrete and continuous random variables are present in a Bayesian network.