Learning Bayesian Network Structure based on Implicit Method
Gang Wang · Journal of Henan Normal University · 2011
In order to overcome the problem that most of the priori knowledge for learning Bayesian network structure is difficulty to obtain,implicit distribution is utilized to improve K2 algorithm.An implicit method score based on implicit network is constructed to replace the score of original algorithm.Experiment results used in standard data sets show that this method without priori knowledge can effectively learn a comparative correct and steady Bayesian network structure,which is as good as the K2 algorithm based on BDe score.