Numerical Characterization of Support Recovery in Sparse Regression with Correlated Design

TL;DR

This paper provides a comprehensive numerical analysis of support recovery in sparse regression with correlated features, highlighting the tradeoffs and conditions affecting feature selection accuracy.

Contribution

It offers the first exhaustive numerical characterization of various estimators and model selection strategies in correlated sparse regression problems.

Findings

SCAD with BIC or empirical Bayes performs best

A fundamental tradeoff exists between false positives and negatives

A transition point exists where selection accuracy degrades

Abstract

Sparse regression is frequently employed in diverse scientific settings as a feature selection method. A pervasive aspect of scientific data that hampers both feature selection and estimation is the presence of strong correlations between predictive features. These fundamental issues are often not appreciated by practitioners, and jeapordize conclusions drawn from estimated models. On the other hand, theoretical results on sparsity-inducing regularized regression such as the Lasso have largely addressed conditions for selection consistency via asymptotics, and disregard the problem of model selection, whereby regularization parameters are chosen. In this numerical study, we address these issues through exhaustive characterization of the performance of several regression estimators, coupled with a range of model selection strategies. These estimators and selection criteria were examined across correlated regression problems with varying degrees of signal to noise, distribution of the non-zero model coefficients, and model sparsity. Our results reveal a fundamental tradeoff between false positive and false negative control in all regression estimators and model selection criteria examined. Additionally, we are able to numerically explore a transition point modulated by the signal-to-noise ratio and spectral properties of the design covariance matrix at which the selection accuracy of all considered algorithms degrades. Overall, we find that SCAD coupled with BIC or empirical Bayes model selection performs the best feature selection across the regression problems considered.