Title :
Matrix Factorization Techniques for Recommender Systems
Author :
Koren, Yehuda ; Bell, Robert ; Volinsky, Chris
Author_Institution :
Yahoo Res., Santa Clara, CA, USA
Abstract :
As the Netflix Prize competition has demonstrated, matrix factorization models are superior to classic nearest neighbor techniques for producing product recommendations, allowing the incorporation of additional information such as implicit feedback, temporal effects, and confidence levels.
Keywords :
information filtering; matrix decomposition; retail data processing; Netflix Prize competition; matrix factorization technique; nearest neighbor technique; product recommendation system; recommender system; Bioinformatics; Collaboration; Filtering; Genomics; Motion pictures; Nearest neighbor searches; Predictive models; Recommender systems; Sea measurements; Computational intelligence; Matrix factorization; Netflix Prize;
DOI :
10.1109/MC.2009.263
