In poker, players tend to play sub-optimally due to the uncertainty in the game. Payoffs can be maximized by exploiting these sub-optimal tendencies. One way of realizing this is to acquire the opponent strategy by recognizing the key patterns in its style of play. Existing studies on opponent modeling in poker aim at predicting opponent's future actions or estimating opponent's hand. In this study, we propose a machine learning method for acquiring the opponent's behavior for the purpose of predicting opponent's future actions. We derived a number of features to be used in modeling opponent's strategy. Then, an ensemble learning method is proposed for generalizing the model. The proposed approach is tested on a set of test scenarios and shown to be effective.
展开▼
