Adaptive Online Opponent Game Policy Modeling with Association Rule Mining

Opponent modeling is a research aspect that needs close attention when facing an opponent in a complex game environment. Real-Time Strategy (RTS) games are a representative of one of the highest complex game environments. In RTS games, the players' tactical and strategical decisions need constant adaptation in order to win the game. In this work, a preliminary study is reported on opponent modeling in RTS games via the incorporation of association rule mining to identify game policies in online mode. Such insight into opponents' mode of operation should provide a player with vital information regarding choosing advantageous gameplay countermeasures against an opponent.