Building Endgame Data set to Improve Opponent Modeling Approach

Opponent modeling is an essential approach for building competitive computer agents in imperfect information games. This paper presents a novel approach to accelerate the convergence process in opponent modeling. The approach applies neural network (ANN) to abstract and build an endgame data set of imperfect information game. Based on a labeled database of author's previous work, several parameters are trained to represent features which decide whether or not a random created endgame is available. The verification experiments is based on a poker game opponent modeling approach which is separately trained on different data set to build K-model clustering opponent models. Based on the mix data set of endgame and normal data set, the opponent modeling approach shows nearly performance while convergence much faster.