Collaborative Filtering

Collaborative filtering operates by analyzing user preferences and items rated/purchased to make recommendations. Using matrix factorization (like Singular Value Decomposition, SVD), we project users and items into a shared lower-dimensional latent space to predict ratings.