Utility-Based Multi-Criteria Recommender Systems

Recommender systems have been demonstrated as a useful tool in assisting decision makings. Multi-criteria recommender systems take advantage user preferences in multiple criteria to produce better recommendations. In this paper, we propose a utility-based multi-criteria recommendation algorithm, in which we learn the user expectations by different learning-to-rank methods. Our experimental results based on the real-world data sets demonstrate the effectiveness of the proposed models.