MCL: Mixed-Centric Loss for Collaborative Filtering.
Zhaolin GaoZhaoyue ChengFelipe PérezJianing SunMaksims VolkovsPublished in: WWW (2022)
Keyphrases
- collaborative filtering
- recommender systems
- matrix factorization
- user ratings
- data sparsity
- recommendation systems
- personalized recommendation
- probabilistic matrix factorization
- information filtering
- user centric
- maximum margin matrix factorization
- demographic information
- latent factor models
- deal with information overload
- information loss
- pearson correlation coefficient
- recommendation quality
- multi aspect
- product recommendation
- cold start problem
- low rank
- website
- error propagation