Generalizing matrix factorization through flexible regression priors
Authors:
Open Access
- 23 October 2011
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 13-20
Abstract
Predicting user "ratings" on items is a crucial task in recommender systems. Matrix factorization methods that computes a low-rank approximation of the incomplete user-item rating matrix provide state-of-the-art performance, especially for users and items with several past ratings (warm starts). However, it is a challenge to generalize such methods to users and items with few or no past ratings (cold starts). Prior work [4][32] have generalized matrix factorization to include both user and item features for performing better regularization of factors as well as provide a model for smooth transition from cold starts to warm starts. However, the features were incorporated via linear regression on factor estimates. In this paper, we generalize this process to allow for arbitrary regression models like decision trees, boosting, LASSO, etc. The key advantage of our approach is the ease of computing --- any new regression procedure can be incorporated by "plugging" in a standard regression routine into a few intermediate steps of our model fitting procedure. With this flexibility, one can leverage a large body of work on regression modeling, variable selection, and model interpretation. We demonstrate the usefulness of this generalization using the MovieLens and Yahoo! Buzz datasets.Keywords
This publication has 17 references indexed in Scilit:
- Regression-based latent factor models2009
- The Elements of Statistical Learning2009
- Bayesian probabilistic matrix factorization using Markov chain Monte Carlo2008
- Naïve filterbots for robust cold-start recommendations2006
- Fast maximum margin matrix factorization for collaborative prediction2005
- Random ForestsMachine Learning, 2001
- Maximizing Generalized Linear Mixed Model Likelihoods With an Automated Monte Carlo EM AlgorithmJournal of the Royal Statistical Society Series B: Statistical Methodology, 1999
- FabCommunications of the ACM, 1997
- Generalized Linear Models1989
- Penalized Likelihood for General Semi-Parametric Regression ModelsInternational Statistical Review, 1987