A case study of Empirical Bayes in User-Movie Recommendation system

TL;DR

This paper explores the application of Empirical Bayes methods to optimize hyperparameters in Bayesian user-movie recommendation systems, demonstrating practical benefits in initial parameter tuning and convergence speed.

Contribution

It formulates Empirical Bayes for hyperparameter tuning in collaborative filtering and shows its effectiveness on the MovieLens dataset.

Findings

Empirical Bayes provides good initial hyperparameter estimates.

It accelerates convergence in MCMC-based models.

Useful for datasets with slow or oscillating MCMC convergence.

Abstract

In this article we provide a formulation of empirical bayes described by Atchade (2011) to tune the hyperparameters of priors used in bayesian set up of collaborative filter. We implement the same in MovieLens small dataset. We see that it can be used to get a good initial choice for the parameters. It can also be used to guess an initial choice for hyper-parameters in grid search procedure even for the datasets where MCMC oscillates around the true value or takes long time to converge.