Discussion Forum: Preference learning for affective modeling

There is an increasing trend towards personalization of services and interaction in computer games, e-commerce, marketing as well as other fields. The use of computational models for learning to predict user emotional preferences is of significant importance towards system personalization.

An alternative experimental approach to affective modeling via preference learning is presented in this mini-tutorial. Preference learning is a machine learning research area that aids in the process of exploiting a set of specific features of an individual in an attempt to predict her preferences. Comparative studies of alternative instance preference learning algorithms will be presented while several computer and physical interactive game case studies will be used to showcase the effectiveness of the methodology.