Which bridge estimator is the best for variable selection? (Q2215760)

From the abstract: ``We study the problem of variable selection for linear models under the high-dimensional asymptotic setting, where the number of observations \(n\) grows at the same rate as the number of predictors \(p\). We consider two-stage variable selection techniques (TVS) in which the first stage uses bridge estimators to obtain an estimate of the regression coefficients, and the second stage simply thresholds this estimate to select the ``important'' predictors. The asymptotic false discovery proportion AFDP and true positive proportion (ATPP) of these TVS are evaluated. We prove that for a fixed ATPP, in order to obtain a smaller AFDP, one should pick a bridge estimator with smaller asymptotic mean square error in the first stage of TVS. Based on such principled discovery, we present a sharp comparison of different TVS, via an in-depth investigation of the estimation properties of bridge estimators. Rather than ``orderwise'' error bounds with loose constants, our analysis focuses on precise error characterization. Various interesting signal-to-noise ratio and sparsity settings are studied. Our results offer new and thorough in-sights into high-dimensional variable selection.'' For instance, the variable selection of Lasso can be improved by debiasing and thresholding; a TVS with ridge in its first stage outperforms TVS with other bridge estimators for large values of noise; the optimality of two-stage Lasso among two-stage bridge estimators holds for very sparse signals until the signal strength is below some threshold. The authors conduct numerical experiments to support their theoretical findings and validate the scope of their main conclusions for general design matrices.