Abstract:Heavy-tailed errors impair the accuracy of the least squares estimate, which can be spoiled by a single grossly outlying observation. As argued in the seminal work of Peter Huber in 1973 [Ann. Statist.1 (1973) 799-821], robust alternatives to the method of least squares are sorely needed. To achieve robustness against heavy-tailed sampling distributions, we revisit the Huber estimator from a new perspective by letting the tuning parameter involved diverge with the sample size. In this paper, we develop nonasymptotic concentration results for such an adaptive Huber estimator, namely, the Huber estimator with the tuning parameter adapted to sample size, dimension, and the variance of the noise. Specifically, we obtain a sub-Gaussian-type deviation inequality and a nonasymptotic Bahadur representation when noise variables only have finite second moments. The nonasymptotic results further yield two conventional normal approximation results that are of independent interest, the Berry-Esseen inequality and Cramér-type moderate deviation. As an important application to large-scale simultaneous inference, we apply these robust normal approximation results to analyze a dependence-adjusted multiple testing procedure for moderately heavy-tailed data. It is shown that the robust dependence-adjusted procedure asymptotically controls the overall false discovery proportion at the nominal level under mild moment conditions. Thorough numerical results on both simulated and real datasets are also provided to back up our theory.

THE ASYMPTOTICALLY OPTIMAL EMPIRICAL BAYES ESTIMATION IN MULTIPLE LINEAR REGRESSION MODEL

The Convergence Rates of Empirical Bayes Estimation in a Multiple Linear Regression Model

ASYMPTOTICALLY OPTIMAL EMPIRICAL BAYES ESTIMATION FOR THE PARAMETERS OF MULTI-PARAMETER DISCRETE EXPONENTIAL FAMILY

On Empirical Bayes Estimation of Multivariate Regression Coefficient.

A Estimate Class of G-M Linear Model

Convergence rates of empirical Bayesian estimation in a class of linear models

Asymptotic Theory of Least Distances Estimate in Multivariate Linear Models

Empirical Bayes Test of Regression Coefficient in a Multiple Linear Regression Model

On Empirical Bayes Estimation of Variance Components in Random Effects Model

The Superiority of Empirical Bayes Estimator of Parameters in Linear Model

EMPIRICAL BAYESIAN ESTIMATION OF THE LOCATION PARAMETER IN DOUBLE EXPONENTIAL DISTRIBUTION

Inference for High-Dimensional Linear Expectile Regression with De-Biasing Method

Asymptotically Optimal Empirical Bayes Estimation for Variance Components in Random Effects Models

A NEW PERSPECTIVE ON ROBUST M-ESTIMATION: FINITE SAMPLE THEORY AND APPLICATIONS TO DEPENDENCE-ADJUSTED MULTIPLE TESTING

A flexible empirical Bayes approach to multiple linear regression and connections with penalized regression

Model Averaging Estimation for Nonparametric Varying-Coefficient Models with Multiplicative Heteroscedasticity

Empirical Bayes Estimation and Its Superiority for Two-Way Classification Model

Asymptotic properties on high-dimensional multivariate regression M-estimation

RATES OF CONVERGENCE OF M-ESTIMATORS FOR PARTLY LINEAR-MODELS INVOLVING TIME-SERIES

Empirical Bayes via ERM and Rademacher complexities: the Poisson model

Nonparametric Regression Estimation with General Parametric Error Covariance: a More Efficient Two-Step Estimator