ProtoMF: Prototype-based Matrix Factorization for Effective and Explainable Recommendations.
Alessandro B. MelchiorreNavid RekabsazChristian GanhörMarkus SchedlPublished in: RecSys (2022)
Keyphrases
- matrix factorization
- recommender systems
- collaborative filtering
- latent factor models
- nonnegative matrix factorization
- factorization methods
- item recommendation
- low rank
- negative matrix factorization
- data sparsity
- cold start problem
- variational bayesian
- missing data
- factor analysis
- personalized ranking
- pairwise
- cold start
- topic models