Towards a more realistic evaluation: testing the ability to predict future tastes of matrix factorization-based recommenders.
Pedro G. CamposFernando DíezManuel A. Sánchez-MontañésPublished in: RecSys (2011)
Keyphrases
- matrix factorization
- collaborative filtering
- recommender systems
- user preferences
- low rank
- information overload
- user ratings
- missing data
- negative matrix factorization
- factor analysis
- nonnegative matrix factorization
- probabilistic matrix factorization
- item recommendation
- data sparsity
- implicit feedback
- factorization methods
- data matrix
- recommendation systems
- stochastic gradient descent
- relevance feedback
- variational bayesian
- user behavior