A reinforcement learning system by using a mixture model of Bayesian network | IEEE Conference Publication | IEEE Xplore