USER PROFILING ON THE BASIS OF INTERACTIONS WITH A COMPUTER GAME

Adapting computer software to the individual user during run-time could offer
substantial advantages over the current practice of tailoring software to groups of
users during the development process (Stewart 2007; Charles et al. 2005; Charles &
Black 2004; Houlette 2004). In order to achieve this, the computer requires
information about the user, yet its ability to perceive them is severely limited
(Suchman 2006 p.167; Fisher 2001). In an effort to address this shortcoming, this
dissertation examines the potential for determining an individual’s personality
through analysis of their interactions with commercial computer games – which, in
common with cinema and literature, work on an underlying model of reality – as well
as their performance in game elements using an underlying general intelligence
factor, and their emotional state from visual and physiological cues. Through a
program of original primary research, it demonstrates that data pertaining to several
of the big five personality factors can be captured from interactions with a
commercial computer game, and explores methods for predicting these personality
traits using regression analysis and clustering techniques. It also employs a series of
factor analyses to investigate the latent variables present in interactions with a
computer role-playing game, as a foundation for further work.