Incorporating Eye-Tracking and Facial Expression Recognition for Preference Prediction in Recommender Systems

Human behavior contains significant information that can greatly contribute to improving the way we interact with computer systems. Getting a better understanding of a user during their interaction with a computer system will greatly improve usability and user satisfaction. Current recommender systems lack such human behavioral indicators. Adding these indicators can be a valuable factor in improving typical recommender systems. This paper proposed a framework that incorporates behavioral indicators in recommender systems to predict user preference. The behavioral indicators chosen for this framework are eye-tracking and emotion recognition through facial expressions.