Using reference models in variable selection
Authors:
Preprint
Abstract
Variable selection, or more generally, model reduction is an important aspect of the statistical workflow aiming to provide insights from data. In this paper, we discuss and demonstrate the benefits of using a reference model in variable selection. A reference model acts as a noise-filter on the target variable by modeling its data generating mechanism. As a result, using the reference model predictions in the model selection procedure reduces the variability and improves stability leading to improved model selection performance. Assuming that a Bayesian reference model describes the true distribution of future data well, the theoretically preferred usage of the reference model is to project its predictive distribution to a reduced model leading to projection predictive variable selection approach. Alternatively, reference models may also be used in an ad-hoc manner in combination with common variable selection methods. In several numerical experiments, we investigate the performance of the projective prediction approach as well as alternative variable selection methods with and without reference models. Our results indicate that the use of reference models generally translates into better and more stable variable selection. Additionally, we demonstrate that the projection predictive approach shows superior performance as compared to alternative variable selection methods independently of whether or not they use reference models.Keywords
All Related Versions
This publication has 14 references indexed in Scilit:
- Visualization in Bayesian WorkflowJournal of the Royal Statistical Society Series A: Statistics in Society, 2019
- Variable selection – A review and recommendations for the practicing statisticianBiometrical Journal, 2018
- Practical Bayesian model evaluation using leave-one-out cross-validation and WAICStatistics and Computing, 2016
- Comparison of Bayesian predictive methods for model selectionStatistics and Computing, 2016
- A survey of Bayesian predictive methods for model assessment, selection and comparisonStatistics Surveys, 2012
- “Preconditioning” for feature selection and regression in high-dimensional problemsThe Annals of Statistics, 2008
- Microarrays, Empirical Bayes and the Two-Groups ModelStatistical Science, 2008
- Prediction by Supervised Principal ComponentsJournal of the American Statistical Association, 2006
- Understanding neural networks using regression trees: an application to multiple myeloma survival dataStatistics in Medicine, 2001
- Regression Shrinkage and Selection Via the LassoJournal of the Royal Statistical Society Series B: Statistical Methodology, 1996