Theoretically Accurate Regularization Technique for Matrix Factorization based Recommender Systems

TL;DR

This paper introduces a theoretically sound regularization method for matrix factorization in recommender systems, addressing the limitations of traditional scalar regularization coefficients and improving accuracy and fairness.

Contribution

It proves the invalidity of common scalar regularization approaches and proposes a new theoretically accurate regularization technique that outperforms existing methods.

Findings

The new method improves recommendation accuracy.

The approach enhances fairness in recommendations.

Traditional scalar regularization is theoretically invalid.

Abstract

Regularization is a popular technique to solve the overfitting problem of machine learning algorithms. Most regularization technique relies on parameter selection of the regularization coefficient. Plug-in method and cross-validation approach are two most common parameter selection approaches for regression methods such as Ridge Regression, Lasso Regression and Kernel Regression. Matrix factorization based recommendation system also has heavy reliance on the regularization technique. Most people select a single scalar value to regularize the user feature vector and item feature vector independently or collectively. In this paper, we prove that such approach of selecting regularization coefficient is invalid, and we provide a theoretically accurate method that outperforms the most widely used approach in both accuracy and fairness metrics.