Maximum-Margin Matrix Factorization

We present a novel approach to collaborative prediction, using low-norm instead of low-rank factorizations. The approach is inspired by, and has strong connections to, large-margin linear discrimination. We show how to learn low-norm factorizations by solving a semi-definite program, and discuss generalization error bounds for them.

• Type: Conference
• Year: 2004
• Where: NIPS
• Authors: Nathan Srebro, Jason D. M. Rennie, Tommi Jaakkola
• Keywords: Generalization Error Bounds, Large-margin Linear Discrimination, Low-rank Factorizations, NIPS 2004, NIPS 2007