Description
Recommender-system experiments can be built and evaluated with collaborative-filtering algorithms. Data scientists and students can load rating datasets, compare models, and measure prediction quality from notebooks or scripts. It is a research and prototyping library, not a hosted recommendation service.