Investigating individual game-play patterns using a self-organzing map

Computer games are played by a diverse range of players which has as diverse preferences and strategies to overcome the game. Most of these strategies are forecasted by the developers and is addressed accordingly in game AI, so the feeling of engagement with the game is not lost. However, with time, these game AI strategies become mundane and repetitive which generally results in exploitation by the players. This could be avoided if game AI is catered towards individual user's preferences and quirks. However, this type of adaptation seems distant with the current game AI methods. One viable approach of achieving this level of personalization is to learn player tactics from the player itself and use it to adapt the game AI to create an absorbing play experience. This paper investigates the possibility of understanding decision making patterns of an individual player using play data from the 2D arcade game Pacman via an unsupervised learning approach.