On the sample covariance matrix estimator of reduced effective rank population matrices, with applications to fPCA

DOI10.3150/14-BEJ602MaRDI QIDQ2348740zbMATH OpenFDO

Authors Florentina Bunea, Luo Xiao

Publication date 15 June 2015

Published in Bernoulli (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1212.5321, https://projecteuclid.org/euclid.bj/1429624975

zbMATH Keywords

minimax ratessample covariance matrixFrobenius normdecaying spectrum

Mathematics Subject Classification ID

Factor analysis and principal components; correspondence analysis (62H25) Estimation in multivariate analysis (62H12)

Abstract: This work provides a unified analysis of the properties of the sample covariance matrix $Sigm{a}_n$ over the class of $pimesp$ population covariance matrices $Sigma$ of reduced effective rank $r_e(Sigma)$. This class includes scaled factor models and covariance matrices with decaying spectrum. We consider $r_e(Sigma)$ as a measure of matrix complexity, and obtain sharp minimax rates on the operator and Frobenius norm of $Sigm{a}_n-Sigma$, as a function of $r_e(Sigma)$ and $|Sigma{|}_2$, the operator norm of $Sigma$. With guidelines offered by the optimal rates, we define classes of matrices of reduced effective rank over which $Sigm{a}_n$ is an accurate estimator. Within the framework of these classes, we perform a detailed finite sample theoretical analysis of the merits and limitations of the empirical scree plot procedure routinely used in PCA. We show that identifying jumps in the empirical spectrum that consistently estimate jumps in the spectrum of $Sigma$ is not necessarily informative for other goals, for instance for the selection of those sample eigenvalues and eigenvectors that are consistent estimates of their population counterparts. The scree plot method can still be used for selecting consistent eigenvalues, for appropriate threshold levels. We provide a threshold construction and also give a rule for checking the consistency of the corresponding sample eigenvectors. We specialize these results and analysis to population covariance matrices with polynomially decaying spectra, and extend it to covariance operators with polynomially decaying spectra. An application to fPCA illustrates how our results can be used in functional data analysis.

Recommendations

• Spectrum estimation for large dimensional covariance matrices using random matrix theory
• Spectrum estimation: a unified framework for covariance matrix estimation and PCA in large dimensions
• Efficient estimation of linear functionals of principal components
• ON ESTIMATION OF THE POPULATION SPECTRAL DISTRIBUTION FROM A HIGH‐DIMENSIONAL SAMPLE COVARIANCE MATRIX
• Finite sample approximation results for principal component analysis: A matrix perturbation approach

Cites work

• scientific article; zbMATH DE number 3886886 (Why is no real title available?)
• Adaptive thresholding for sparse covariance matrix estimation
• Arbitrage, Factor Structure, and Mean-Variance Analysis on Large Asset Markets
• Asymptotic Theory for Principal Component Analysis
• Asymptotic theory for the principal component analysis of a vector random function: Some applications to statistical inference
• Common functional principal components
• Covariance regularization by thresholding
• Determining the Number of Factors in Approximate Factor Models
• Eigenvalues of large sample covariance matrices of spiked population models
• Factor models and variable selection in high-dimensional regression analysis
• Finite sample approximation results for principal component analysis: A matrix perturbation approach
• Functional Data Analysis for Sparse Longitudinal Data
• High-dimensional covariance matrix estimation with missing observations
• Inference for Density Families Using Functional Principal Component Analysis
• Matrix Analysis
• On Properties of Functional Principal Components Analysis
• On the distribution of the largest eigenvalue in principal components analysis
• On the sample covariance matrix estimator of reduced effective rank population matrices, with applications to fPCA
• Optimal rates of convergence for covariance matrix estimation
• Optimal selection of reduced rank estimators of high-dimensional matrices
• Properties of principal component methods for functional and longitudinal data analysis
• Regularized estimation of large covariance matrices
• User-friendly tail bounds for sums of random matrices

Cited in

(27)

• Concentration inequalities for statistical inference
• Improved second order estimation in the singular multivariate normal model
• A joint design for functional data with application to scheduling ultrasound scans
• Fast covariance estimation for high-dimensional functional data
• New asymptotic results in principal component analysis
• Adaptive estimation of the copula correlation matrix for semiparametric elliptical copulas
• Dimension-free bounds for sums of dependent matrices and operators with heavy-tailed distributions
• scientific article; zbMATH DE number 7415120 (Why is no real title available?)
• Inference in latent factor regression with clusterable features
• Non-asymptotic properties of spectral decomposition of large Gram-type matrices and applications
• Posterior contraction in sparse Bayesian factor models for massive covariance matrices
• On cumulative slicing estimation for high dimensional data
• Bootstrapping the operator norm in high dimensions: error estimation for covariance matrices and sketching
• Compressed covariance estimation with automated dimension learning
• On the sample covariance matrix estimator of reduced effective rank population matrices, with applications to fPCA
• Optimal estimation and rank detection for sparse spiked covariance matrices
• Representation and reconstruction of covariance operators in linear inverse problems
• Bagging-enhanced sampling schedule for functional quadratic regression
• Moment bounds for large autocovariance matrices under dependence
• Bootstrapping max statistics in high dimensions: near-parametric rates under weak variance decay and application to functional and multinomial data
• Blind Identification of Stochastic Block Models from Dynamical Observations
• Estimation of low-rank covariance function
• Non-asymptotic rate for high-dimensional covariance estimation with non-independent missing observations
• Sparse and integrative principal component analysis for multiview data
• Detecting approximate replicate components of a high-dimensional random vector with latent structure
• ECA: High-Dimensional Elliptical Component Analysis in Non-Gaussian Distributions
• Rates of Bootstrap Approximation for Eigenvalues in High-Dimensional PCA

This page was built for publication: On the sample covariance matrix estimator of reduced effective rank population matrices, with applications to fPCA

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2348740)