MF
Approximate latent representation factors used for pattern learning in machine learning while Matrix Decomposition extracts structured components. It is essentially same as Representation Learning
Data sparsity 낮은 heavy user에게 Serendipity 있는 추천 해주는 것에 적합하다
Expressing user-item interaction in the form of inner product (linear)
- Generalized Matrix Factorization
- Neural matrix factorization
- MLP-based Factorization