Improving rating estimation in recommender systems using aggregation- and variance-based hierarchical models

ABSTRACT

Previous work on using external aggregate rating information showed that this information can be incorporated in several different types of recommender systems and improves their performance. In this paper, we propose a more general class of methods that combine external aggregate information with individual ratings in a novel way. Unlike the previously proposed methods, one of the defining features of this approach is that it takes into the consideration not only the aggregate average ratings but also the variance of the aggregate distribution of ratings. The methods proposed in this paper estimate unknown ratings by finding an optimal linear combination of individual-level and aggregate-level rating estimators in a form of a hierarchical regression (HR) model that is grounded in the theory of statistics and machine learning.

The proposed HR model is general enough so that the standard individual-level recommender systems and naive aggregate methods constitute special cases of this model.

We show that for the general HR model, the presence of the aggregate variance, surprisingly, does not significantly improve estimation of unknown ratings vis-a-vis the case when only aggregate average ratings are considered.

In the paper, we experimentally show that the optimal linear combination approach significantly dominates all other special cases, including the classical non-aggregated case and our previously studied aggregate methods, and therefore is the method of choice.

REFERENCES

• 1: G. Adomavicius and Y. Kwon. Overcoming Accuracy-Diversity Tradeoff in Recommender Systems: A Variance-Based Approach. Proceedings of WITS'08, Paris, France, 2008.
• 2: G. Adomavicius, R. Sankaranarayanan, S. Sen, and A. Tuzhilin. Incorporating contextual information in recommender systems using a multidimensional approach. ACM Transactions on Information Systems (TOIS), 23(1):103--145, 2005.
• 3: G. Adomavicius and A. Tuzhilin. Multidimensional recommender systems: a data warehousing approach. In 2nd Intl. Workshop on Electronic Commerce. LNCS 2232, 2001.
• 4: D. Agarwal, A. Broder, D. Chakrabarti, D. Diklic, V. Josifovski, and M. Sayyadian. Estimating rates of rare events at multiple resolutions. Proceedings of the 13th ACM SIGKDD, pages 16--25, 2007.
• 5: A. Ansari, S. Essegaier, and R. Kohli. Internet Recommendation Systems. J. of Marketing Research, 37(3), 2000.
• 6: R. Bell, Y. Koren, and C. Volinsky. Modeling relationships at multiple scales to improve accuracy of large recommender systems. Procs of the 13th ACM SIGKDD, 2007.
• 7: J. Bennett and S. Lanning. The Netflix Prize. Proceedings of KDD Cup and Workshop 2007, 2007.
• 8: C. Bishop and N. Nasrabadi. Pattern Recognition and Machine Learning. J. of Electronic Imaging, 16:049901, 2007.
• 9: J. Bollen. Group user models for personalized hyperlink recommendations. Procs of the Int. Conference on Adaptive Hypermedia and Adaptive Web--Based Systems, 2000.
• 10: IMDB. http://www.imdb.com. 2006.
• 11: A. Jameson and B. Smyth. Recommendation to Groups, chapter The Adaptive Web: Methods and strategies of web personalization. Springer, 2006.
• 12: MovieLens. available at http://www.grouplens.org/node/73, 2006.
• 13: M. O'Connor, D. Cosley, J. A. Konstan, and J. Riedl. Polylens: A recommender system for groups of users. Procs of ECSCW, 2001.
• 14: S. W. Raudenbush and A. S. Bryk. Hierarchical Linear Models: Applications and Data Analysis Methods. Sage Publications, Inc, 2001.
• 15: B. Sarwar, G. Karypis, J. Konstan, and J. Reidl. Item--based collaborative filtering recommendation algorithms. Procs of the 10th int. conference on World Wide Web, 2001.
• 16: A. Umyarov and A. Tuzhilin. Leveraging aggregate ratings for better recommendations. Procs of the 2007 ACM Recommender systems, 2007.
• 17: A. Umyarov and A. Tuzhilin. Improving Collaborative Filtering Recommendations Using External Data. Proceedings of the IEEE ICDM 2008 Conference, 2008.
• 18: A. Umyarov and A. Tuzhilin. Leveraging aggregate ratings for improving predictive performance of recommender systems. Working paper. Stern School of Business. New York University. CeDER-08-03, 2008

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.1 MODELS AND PRINCIPLES
      H.1.2 User/Machine Systems
          Subjects: Human information processing

General Terms:
Algorithms, Design, Theory

Keywords:
aggregate ratings, aggregate variance, hierarchical models, predictive models