What about Interpreting Features in Matrix Factorization-based Recommender Systems as Users?
Marharyta AleksandrovaArmelle BrunAnne BoyerOleg ChertovPublished in: HT (Doctoral Consortium / Late-breaking Results / Workshops) (2014)
Keyphrases
- recommender systems
- matrix factorization
- collaborative filtering
- data sparsity
- cold start problem
- implicit feedback
- user preferences
- low rank
- user model
- factorization methods
- user profiles
- cold start
- recommendation systems
- information overload
- latent factors
- recommendation quality
- negative matrix factorization
- feature vectors
- item recommendation
- image features
- variational bayesian
- stochastic gradient descent
- feature extraction
- user ratings
- personalized recommendation
- nonnegative matrix factorization
- co occurrence
- factor analysis
- user feedback
- personalized ranking
- probabilistic matrix factorization
- user interests
- user behavior
- feature space