Reinforcement learning method based on semi-parametric regression model | IEEE Conference Publication | IEEE Xplore