Investigation of Various Matrix Factorization Methods for Large Recommender Systems.
Gábor TakácsIstván PilászyBottyán NémethDomonkos TikkPublished in: ICDM Workshops (2008)
Keyphrases
- factorization methods
- matrix factorization
- recommender systems
- collaborative filtering
- low rank
- negative matrix factorization
- factor analysis
- nonnegative matrix factorization
- factorization method
- user preferences
- user profiles
- implicit feedback
- tensor factorization
- user model
- personalized recommendation
- pairwise
- latent factors
- user feedback
- user interests
- data analysis
- missing data
- low dimensional
- search engine
- linear combination