Matrix Factorization
A model-based collaborative filtering method that decomposes a high-dimensional sparse rating matrix into a user latent-vector matrix and an item latent-vector matrix, predicting ratings via their inner product. The technique was brought into industry mainstream through the Netflix Prize competition.
Timeline (last 90 days)
Matrix factorization only models the two-dimensional user-item relationship and cannot natively incorporate context such as time, device, and session, so it requires context-extended versions or combination with feature models (such as GBDT).
Model-based collaborative filtering uses matrix factorization to decompose a high-dimensional sparse rating matrix into two low-dimensional dense matrices representing user latent vectors and item latent vectors, and then predicts ratings via inner product.
Matrix factorization decomposes a user-item interaction matrix into the product of two low-rank matrices, capturing latent features of users and items in a latent space.
All Facts (3)
Matrix factorization decomposes a user-item interaction matrix into the product of two low-rank matrices, capturing latent features of users and items in a latent space.
60%UnverifiedMatrix factorization only models the two-dimensional user-item relationship and cannot natively incorporate context such as time, device, and session, so it requires context-extended versions or combination with feature models (such as GBDT).
50%UnverifiedModel-based collaborative filtering uses matrix factorization to decompose a high-dimensional sparse rating matrix into two low-dimensional dense matrices representing user latent vectors and item latent vectors, and then predicts ratings via inner product.
50%