Model-based CF

Approximate latent representation factors used for pattern learning in machine learning while

Matrix Decomposition extracts structured components. It is essentially same as

Representation Learning