Learning User Preference Models under Uncertainty for Personalized Recommendation

Preference modeling has a crucial role in customer relationship management systems. Traditional approaches to preference modeling are based on decision and utility theory by explicitly querying users about the behavior of value function, or utility of every outcome with regard to each decision criterion. They are error-prone and labor intensive. To address these limitations, computer based implicit elicitation approaches have been proposed. However, the extant approaches to implicit elicitation in preference modeling have failed to: (i) integrate user feedbacks and item attributes; (ii) take into account of the subjective, incomplete, imprecise, and vague nature of features of an item, and features of a user preference; (iii) quantify how much a user likes, dislikes, or be indifferent to a given item; and (iv) provide a complete preference model. We propose a novel knowledge representation method for item and user preference that accounts for uncertainty due to the subjectivity, vagueness and imprecision using concepts from the fuzzy set and logic theory. A comprehensive preference model that accounts for positive, negative, neutral and in-deterministic categories of user preferences is defined. Furthermore, algorithms are developed for learning user preferences, and for prediction and recommendation. An evaluation with a benchmark dataset on movies shows that the accuracy in predicting user preference is found to be nearly twice that of random prediction. Additionally, the proposed approach outperformed the state-of-the-art approaches in terms of precision, recall, and F1-measure. The findings of this study have significant implications for preference modeling, recommender systems and customer relationships management systems.