Context-Based Satisfaction Modelling for Personalized Recommendations

Approaches for the personalized recommendations focus mainly on the user´s activity over various portals. User´s preferences are not dependent on the long term users´ history only, but actual user´s situation plays crucial role in the user´s preferences adjustment and formation. An item liked by the user in some context, can be disliked by the same user in the other context. For the considering this users´ variability we propose a novel approach for the user´s satisfaction modelling based on incorporating the user´s context into the rating prediction and consideration of previous users´ rating history. Our novel approach reflects natural characteristics of user´s context, when the various context´s settings can influence another context. Proposed approach brings statistically significant improvement in the rating prediction process, thus it can increase user satisfaction during one-session recommendation.