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.

Tail approximations for samples from a finite population with applications to permutation tests

Approximation of the Tail Probability of Randomly Weighted Sums and Applications

Approximation of the Tail Probability of Dependent Random Sums under Consistent Variation and Applications

Approximation of Sums of Locally Dependent Random Variables via Perturbation of Stein Operator

Non-parametric Tests for the Tail Equivalence Via Empirical Likelihood

Empirical Likelihood Based Confidence Intervals For The Tail Index When Gamma <-1/2

Empirical Likelihood Based Inference for Conditional Pareto-type Tail Index

Saddlepoint approximations for likelihood ratio like statistics with applications to permutation tests

The saddlepoint approximation for averages of conditionally independent random variables

Permutation Tests at Nonparametric Rates

Nearly Tight Universal Bounds for the Binomial Tail Probabilities

Multivariate saddlepoint approximations in tail probability and conditional inference

Fast Approximation of Small P-values in Permutation Tests by Partitioning the Permutations

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

Some Approximations to the Binomial Distribution Function

Hypergeometric Distribution Revisited: Tail Inequalities, Confidence Bounds and Sample Sizes

A simple wide range approximation of symmetric binomial distributions

Tail bounds for empirically standardized sums

Saddlepoint Approximations for the Difference of Order Statistics and Studentized Sample Quantiles

A studentized permutation test in group sequential designs

Nonnormal small jump approximation of infinitely divisible distributions