Non-myopic active learning for recommender systems based on Matrix Factorization.
Rasoul KarimiChristoph FreudenthalerAlexandros NanopoulosLars Schmidt-ThiemePublished in: IRI (2011)
Keyphrases
- matrix factorization
- recommender systems
- active learning
- collaborative filtering
- low rank
- factorization methods
- user preferences
- factor analysis
- cold start problem
- implicit feedback
- learning algorithm
- negative matrix factorization
- nonnegative matrix factorization
- data sparsity
- semi supervised
- machine learning
- random sampling
- supervised learning
- training set
- variational bayesian
- cold start
- transfer learning
- missing data
- rating prediction
- personalized recommendation
- tensor factorization
- relevance feedback
- item recommendation
- personalized ranking
- user interests
- stochastic gradient descent
- user profiles
- labeled data
- user ratings
- user feedback
- recommendation quality
- linear combination
- image classification
- low rank matrix factorization