Cornac is simple, handy and very intuitive. It is designed from the ground-up to faithfully reflect the standard steps taken by researchers to implement and evaluate personalized recommendation models. Only few lines of code are needed for training and benchmarking algorithms.
Development of Cornac is undertaken by experts in machine learning and recommender systems, who harness their knowledge to provide an environment for fair evaluations and comparisons. All the core and key features of Cornac are deeply checked and covered by tests.
Cornac is designed with data sparsity in mind. All computations involving users and items are curried out over observed preferences only. Efficient vector operations such as those supported by Numpy are leveraged as much as possible. Your experiments can easily scale to large datasets.