Leveraging Aggregate Ratings for Better Recommendations
7 Pages Posted: 9 Oct 2008
Date Written: September 2007
The paper presents a method that uses aggregate ratingsprovided by various segments of users for various categoriesof items to derive better estimations of unknown individualratings. This is achieved by converting the aggregate ratingsinto constraints on the parameters of a rating estimationmodel presented in the paper. The paper also demonstratestheoretically that these additional constraints reduce ratingestimation errors resulting in better rating predictions.
Keywords: Recommender systems, Hierarchical Bayesian models, predictive models, aggregate ratings, OLAP
Suggested Citation: Suggested Citation