Gaussian approximations and multiplier bootstrap for maxima of sums of high-dimensional random vectors

Abstract

We derive a Gaussian approximation result for the maximum of a sum of high-dimensional random vectors. Specifically, we establish conditions under which the distribution of the maximum is approximated by that of the maximum of a sum of the Gaussian random vectors with the same covariance matrices as the original vectors. This result applies when the dimension of random vectors ($p$) is large compared to the sample size ($n$); in fact, $p$ can be much larger than $n$, without restricting correlations of the coordinates of these vectors. We also show that the distribution of the maximum of a sum of the random vectors with unknown covariance matrices can be consistently estimated by the distribution of the maximum of a sum of the conditional Gaussian random vectors obtained by multiplying the original vectors with i.i.d. Gaussian multipliers. This is the Gaussian multiplier (or wild) bootstrap procedure. Here too, $p$ can be large or even much larger than $n$. These distributional approximations, either Gaussian or conditional Gaussian, yield a high-quality approximation to the distribution of the original maximum, often with approximation error decreasing polynomially in the sample size, and hence are of interest in many applications. We demonstrate how our Gaussian approximations and the multiplier bootstrap can be used for modern high-dimensional estimation, multiple hypothesis testing, and adaptive specification testing. All these results contain nonasymptotic bounds on approximation errors.

A minor typo has been corrected (last line, page 22, where \max_{j \in w} was missing)

References in corpus (5)

• The Dantzig selector and sparsity oracle inequalities
• Confidence bands in density estimation
• A simple invariance theorem
• Testing Regression Monotonicity in Econometric Models
• To how many simultaneous hypothesis tests can normal, Student's t or bootstrap calibration be applied?

Cited by in corpus (35)

• Gaussian approximation of suprema of empirical processes
• Valid Post-Selection and Post-Regularization Inference: An Elementary, General Approach
• The Hardness of Conditional Independence Testing and the Generalised Covariance Measure
• Gaussian and bootstrap approximations for high-dimensional U-statistics and their applications
• Bootstrap confidence sets under model misspecification
• Comparing large covariance matrices under weak conditions on the dependence structure and its application to gene clustering
• Nonparametric inference in generalized functional linear models
• Change-point detection in high-dimensional covariance structure
• A multivariate Berry--Esseen theorem with explicit constants
• Testing for high-dimensional white noise using maximum cross-correlations
• Simulation-Based Hypothesis Testing of High Dimensional Means Under Covariance Heterogeneity
• Testing Regression Monotonicity in Econometric Models
• DoubleML -- An Object-Oriented Implementation of Double Machine Learning in R
• Confidence regions for entries of a large precision matrix
• Are Discoveries Spurious? Distributions of Maximum Spurious Correlations and Their Applications
• Approximating high-dimensional infinite-order $U$-statistics: statistical and computational guarantees
• Finite sample change point inference and identification for high-dimensional mean vectors
• Causal mediation analysis for stochastic interventions
• Central limit theorems for high dimensional dependent data
• A Simple Adjustment for Bandwidth Snooping
• Lagrangian Inference for Ranking Problems
• Testing the martingale difference hypothesis in high dimension
• On the Post Selection Inference constant under Restricted Isometry Properties
• A Max-Correlation White Noise Test for Weakly Dependent Time Series
• Notes on the dimension dependence in high-dimensional central limit theorems for hyperrectangles
• Nonclassical Berry-Esseen inequalities and accuracy of the bootstrap
• Edge differentially private estimation in the $\beta$-model via jittering and method of moments
• New Edgeworth-type expansions with finite sample guarantees
• On frequentist coverage errors of Bayesian credible sets in moderately high dimensions
• A Berry-Esseen Bound for Vector-valued Martingales
• Statistical Inference For Noisy Matrix Completion Incorporating Auxiliary Information
• An approximate randomization test for high-dimensional two-sample Behrens-Fisher problem under arbitrary covariances
• Central Limit Theorem and Near classical Berry-Esseen rate for self normalized sums in high dimensions
• Some New Asymptotic Theory for Least Squares Series: Pointwise and Uniform Results
• Continuous-Variable Quantum Key Distribution with Composable Security and Tight Error Correction Bound towards Constrained-Device Implementations