Leveraging Aggregate Ratings for Improving Predictive Performance of Recommender Systems

This paper describes an approach for incorporating externally specified aggregate ratings informationinto certain types of recommender systems, including two types of collaborating filteringand a hierarchical linear regression model. First, we present a framework for incorporating aggregaterating information and apply this framework to the aforementioned individual rating models.Then we formally show that this additional aggregate rating information provides more accuraterecommendations of individual items to individual users. Further, we experimentally confirm thistheoretical finding by demonstrating on several datasets that the aggregate rating informationindeed leads to better predictions of unknown ratings. We also propose scalable methods forincorporating this aggregate information and test our approaches on large datasets. Finally, wedemonstrate that the aggregate rating information can also be used as a solution to the cold startproblem of recommender systems