Latent Factor Models and Matrix Factorizations

Definition

Latent Factor models are a state of the art methodology for model-based collaborative filtering. The basic assumption is that there exist an unknown low-dimensional representation of users and items where user-item affinity can be modeled accurately. For example, the rating that a user gives to a movie might be assumed to depend on few implicit factors such as the user’s taste across various movie genres. Matrix factorization techniques are a class of widely successful Latent Factor models that attempt to find weighted low-rank approximations to the user-item matrix, where weights are used to hold out missing entries. There is a large family of matrix factorization models based on choice of loss function to measure approximation quality, regularization terms to avoid overfitting, and other domain-dependent formulations.