A user model for recommendation based on facial expression recognition

Abstract

User modeling is crucial in recommendation system. By analyzing users' behavior, gathering users' interest information, the user model can well express the users' need. While users' emotion is also good expression of users' needs. A user model based on facial expression recognition is built. The user model is on the base of traditional user model, affective function α is provided to reflect user's emotion. As well as the user model updating times t is also discussed in the model. And then a discussion of the updating strategy and application in recommendation is had. © 2011 Springer-Verlag Berlin Heidelberg.