29
1

Error estimation and adaptive tuning for unregularized robust M-estimator

Abstract

We consider unregularized robust M-estimators for linear models under Gaussian design and heavy-tailed noise, in the proportional asymptotics regime where the sample size n and the number of features p are both increasing such that p/nγ(0,1)p/n \to \gamma\in (0,1). An estimator of the out-of-sample error of a robust M-estimator is analysed and proved to be consistent for a large family of loss functions that includes the Huber loss. As an application of this result, we propose an adaptive tuning procedure of the scale parameter λ>0\lambda>0 of a given loss function ρ\rho: choosingλ^\hat \lambda in a given interval II that minimizes the out-of-sample error estimate of the M-estimator constructed with loss ρλ()=λ2ρ(/λ)\rho_\lambda(\cdot) = \lambda^2 \rho(\cdot/\lambda) leads to the optimal out-of-sample error over II. The proof relies on a smoothing argument: the unregularized M-estimation objective function is perturbed, or smoothed, with a Ridge penalty that vanishes as n+n\to+\infty, and show that the unregularized M-estimator of interest inherits properties of its smoothed version.

View on arXiv
Comments on this paper