ABSTRACT
Hybrid recommender systems --- systems using multiple algorithms together to improve recommendation quality --- have been well-known for many years and have shown good performance in recent demonstrations such as the NetFlix Prize. Modern hybridization techniques, such as feature-weighted linear stacking, take advantage of the hypothesis that the relative performance of recommenders varies by circumstance and attempt to optimize each item score to maximize the strengths of the component recommenders. Less attention, however, has been paid to understanding what these strengths and failure modes are. Understanding what causes particular recommenders to fail will facilitate better selection of the component recommenders for future hybrid systems and a better understanding of how individual recommender personalities can be harnessed to improve the recommender user experience. We present an analysis of the predictions made by several well-known recommender algorithms on the MovieLens 10M data set, showing that for many cases in which one algorithm fails, there is another that will correctly predict the rating.
- R. Burke. Hybrid recommender systems: Survey and experiments. User Modeling and User-Adapted Interaction, 12(4):331--370, Nov. 2002. Google ScholarDigital Library
- M. D. Ekstrand, P. Kannan, J. A. Stemper, J. T. Butler, J. A. Konstan, and J. T. Riedl. Automatically building research reading lists. In RecSys '10, pages 159--166. ACM, 2010. Google ScholarDigital Library
- M. D. Ekstrand, M. Ludwig, J. A. Konstan, and J. T. Riedl. Rethinking the recommender research ecosystem: reproducibility, openness, and LensKit. In RecSys '11, pages 133--140. ACM, 2011. Google ScholarDigital Library
- S. Funk. Netflix update: Try this at home. http://sifter.org/~simon/journal/20061211.html, Dec. 2006.Google Scholar
- A. Gunawardana and G. Shani. A survey of accuracy evaluation metrics of recommendation tasks. J. Mach. Learn. Res., 10:2935--2962, 2009. Google ScholarDigital Library
- J. Herlocker, J. A. Konstan, and J. Riedl. An empirical analysis of design choices in Neighborhood-Based collaborative filtering algorithms. Inf. Retr., 5(4):287--310, 2002. Google ScholarDigital Library
- S. M. McNee, J. Riedl, and J. A. Konstan. Making recommendations better: an analytic model for human-recommender interaction. In CHI '06 Extended Abstracts, pages 1103--1108. ACM, 2006. Google ScholarDigital Library
- A. Paterek. Improving regularized singular value decomposition for collaborative filtering. In KDD Cup and Workshop 2007, Aug. 2007.Google Scholar
- P. Resnick, N. Iacovou, M. Suchak, P. Bergstrom, and J. Riedl. GroupLens: an open architecture for collaborative filtering of netnews. In ACM CSCW '94, pages 175--186. ACM, 1994. Google ScholarDigital Library
- B. Sarwar, G. Karypis, J. Konstan, and J. Reidl. Item-based collaborative filtering recommendation algorithms. In ACM WWW '01, pages 285--295. ACM, 2001. Google ScholarDigital Library
- J. Sill, G. Takacs, L. Mackey, and D. Lin. Feature-Weighted linear stacking. arXiv:0911.0460, Nov. 2009.Google Scholar
- R. Torres, S. M. McNee, M. Abel, J. A. Konstan, and J. Riedl. Enhancing digital libraries with TechLensGoogle Scholar
- . In ACM/IEEE JCDL '04, pages 228--236. ACM, 2004.Google Scholar
- A. Umyarov and A. Tuzhilin. Improving rating estimation in recommender systems using aggregation- and variance-based hierarchical models. In RecSys '09, pages 37--44, New York, New York, USA, 2009. ACM. Google ScholarDigital Library
Index Terms
- When recommenders fail: predicting recommender failure for algorithm selection and combination
Recommendations
An empirical comparison of social, collaborative filtering, and hybrid recommenders
Special section on twitter and microblogging services, social recommender systems, and CAMRa2010: Movie recommendation in contextIn the Social Web, a number of diverse recommendation approaches have been proposed to exploit the user generated contents available in the Web, such as rating, tagging, and social networking information. In general, these approaches naturally require ...
Improving Collaborative Filtering Based Recommenders Using Topic Modelling
WI-IAT '14: Proceedings of the 2014 IEEE/WIC/ACM International Joint Conferences on Web Intelligence (WI) and Intelligent Agent Technologies (IAT) - Volume 01Standard Collaborative Filtering (CF) algorithms make use of interactions between users and items in the form of implicit or explicit ratings alone for generating recommendations. Similarity among users or items is calculated purely based on rating ...
Are All Rejected Recommendations Equally Bad?: Towards Analysing Rejected Recommendations
UMAP '19: Proceedings of the 27th ACM Conference on User Modeling, Adaptation and PersonalizationWhen evaluating algorithms that recommend a list of relevant items to a user, it is common to use metrics such as precision to measure the system accuracy. When computing precision, one computes the number of items that were selected by the user among ...
Comments