Java Deep Learning Projects

Factorization machines in recommender systems

In real life, most recommendation problems assume that we have a rating dataset formed by a collection of (user, item, and rating) tuples. However, in many applications, we have plenty of item metadata (tags, categories, and genres) that can be used to make better predictions.

This is one of the benefits of using FMs with feature-rich datasets, because there is a natural way in which extra features can be included in the model, and higher-order interactions can be modeled using the dimensionality parameter.

A few recent types of research show which feature-rich datasets give better predictions:

• Xiangnan He and Tat-Seng Chua, Neural Factorization Machines for Sparse Predictive Analytics. During ...