tlooto
Sign up
Computer ScienceMathematics

Matrix Factorization Techniques for Recommender Systems

Y. Koren, Robert M. Bell, C. Volinsky

2009.8.1COMPUTER

DOI: 10.1109/mc.2009.263

tlooto Summary

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.

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.

Citation format

KOREN, Y.; BELL, Robert M.; VOLINSKY, C. Matrix factorization techniques for recommender systems. COMPUTER, 2009, 42.