Selective Forgetting for Incremental Matrix Factorization in Recommender Systems.
Pawel MatuszykMyra SpiliopoulouPublished in: Discovery Science (2014)
Keyphrases
- matrix factorization
- recommender systems
- incremental learning
- collaborative filtering
- low rank
- data sparsity
- nonnegative matrix factorization
- factor analysis
- negative matrix factorization
- cold start problem
- probabilistic matrix factorization
- variational bayesian
- data matrix
- implicit feedback
- rating prediction
- personalized recommendation
- user preferences
- user profiles
- factorization methods
- missing data
- multiscale
- stochastic gradient descent
- item recommendation
- recommendation quality
- recommendation algorithms
- sparsity constraints
- tensor factorization
- factorization method
- information overload
- least squares
- pairwise