Rates of Bootstrap Approximation for Eigenvalues in High-Dimensional PCA

DOI10.5705/SS.202021.0158MaRDI QIDQ6069877zbMATH OpenWikidataFDO

Authors Junwen Yao, Miles E. Lopes

Publication date 17 November 2023

Published in STATISTICA SINICA (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2104.07328

zbMATH Keywords

bootstrapcovariance matriceshigh-dimensional statisticsprincipal components analysis

Mathematics Subject Classification ID

Statistics (62-XX)

Abstract: In the context of principal components analysis (PCA), the bootstrap is commonly applied to solve a variety of inference problems, such as constructing confidence intervals for the eigenvalues of the population covariance matrix $Sigma$. However, when the data are high-dimensional, there are relatively few theoretical guarantees that quantify the performance of the bootstrap. Our aim in this paper is to analyze how well the bootstrap can approximate the joint distribution of the leading eigenvalues of the sample covariance matrix $hatSigma$, and we establish non-asymptotic rates of approximation with respect to the multivariate Kolmogorov metric. Under certain assumptions, we show that the bootstrap can achieve the dimension-free rate of $tr(Sigma)/sqrtn$ up to logarithmic factors, where $tr(Sigma)$ is the effective rank of $Sigma$, and $n$ is the sample size. From a methodological standpoint, our work also illustrates that applying a transformation to the eigenvalues of $hatSigma$ before bootstrapping is an important consideration in high-dimensional settings.

Cites work

• Bootstrap confidence intervals. With comments and a rejoinder by the authors
• Bootstrap confidence sets for spectral projectors of sample covariance
• Bootstrap tests and confidence regions for functions of a covariance matrix
• Bootstrapping max statistics in high dimensions: near-parametric rates under weak variance decay and application to functional and multinomial data
• Bootstrapping spectral statistics in high dimensions
• Bootstrapping the operator norm in high dimensions: error estimation for covariance matrices and sketching
• Correction to: Bootstrap tests and confidence regions for functions of a covariance matrix
• Efficient estimation of linear functionals of principal components
• High-dimensional asymptotics of prediction: ridge regression and classification
• High-dimensional covariance matrix estimation with missing observations
• High-Dimensional MANOVA Via Bootstrapping and Its Application to Functional and Sparse Count Data
• High-dimensional probability. An introduction with applications in data science
• scientific article; zbMATH DE number 1104922 (Why is no real title available?)
• scientific article; zbMATH DE number 1964693 (Why is no real title available?)
• Nonlinear shrinkage estimation of large-dimensional covariance matrices
• Normal approximation and concentration of spectral projectors of sample covariance
• Normalizing transformations and bootstrap confidence intervals
• On Gaussian comparison inequality and its application to spectral analysis of large random matrices
• On nondifferentiable functions and the bootstrap
• On parameter transformations and interval estimation
• On the number of principal components in high dimensions
• On the sample covariance matrix estimator of reduced effective rank population matrices, with applications to fPCA
• On Wielandt's inequality and its application to the asymptotic distribution of the eigenvalues of a random symmetric matrix
• Principal component analysis.
• Random design analysis of ridge regression
• Random matrix methods for wireless communications.
• Robust multivariate analysis
• Spectral analysis of large dimensional random matrices
• Statistics and Data Analysis for Financial Engineering
• Tie-respecting bootstrap methods for estimating distributions of sets and functions of eigenvalues
• Variance stabilization and the bootstrap

Cited in

(3)

• A bootstrap method for spectral statistics in high-dimensional elliptical models
• Quantitative limit theorems and bootstrap approximations for empirical spectral projectors
• Improved rates of bootstrap approximation for the operator norm: a coordinate-free approach

This page was built for publication: Rates of Bootstrap Approximation for Eigenvalues in High-Dimensional PCA

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6069877)