Sharp minimax tests for large Toeplitz covariance matrices with repeated observations

DOI10.1016/J.JMVA.2015.09.003zbMATH Open1334.62075arXiv1506.01557OpenAlexW1597752104MaRDI QIDQ268748FDOQ268748

Authors: Cristina Butucea, Rania Zgheib 

Publication date: 15 April 2016

Published in: Journal of Multivariate Analysis (Search for Journal in Brave)

Abstract: We observe a sample of $n$ independent $p$-dimensional Gaussian vectors with Toeplitz covariance matrix $Sigma=[sigm{a}_{|i-j|}{]}_{1leqi,jleqp}$ and $sigm{a}_0=1$. We consider the problem of testing the hypothesis that $Sigma$ is the identity matrix asymptotically when $noinfty$ and $poinfty$. We suppose that the covariances $sigm{a}_k$ decrease either polynomially ($su{m}_{kgeq1}{k}^{2alpha}sigm{a}_k^{2}leqL$ for $alpha>1/4$ and $L>0$) or exponentially ($su{m}_{kgeq1}{e}^{2Ak}sigm{a}_k^{2}leqL$ for $A,L>0$). We consider a test procedure based on a weighted U-statistic of order 2, with optimal weights chosen as solution of an extremal problem. We give the asymptotic normality of the test statistic under the null hypothesis for fixed $n$ and $po+infty$ and the asymptotic behavior of the type I error probability of our test procedure. We also show that the maximal type II error probability, either tend to $0$, or is bounded from above. In the latter case, the upper bound is given using the asymptotic normality of our test statistic under alternatives close to the separation boundary. Our assumptions imply mild conditions: $n=o(p^{2alpha-1/2})$ (in the polynomial case), $n=o(e^p)$ (in the exponential case). We prove both rate optimality and sharp optimality of our results, for $alpha>1$ in the polynomial case and for any $A>0$ in the exponential case. A simulation study illustrates the good behavior of our procedure, in particular for small $n$, large $p$.

Full work available at URL: https://arxiv.org/abs/1506.01557

Recommendations

• Sharp minimax tests for large covariance matrices and adaptation
• Fast nonasymptotic testing and support recovery for large sparse Toeplitz covariance matrices
• Adaptive test for large covariance matrices with missing observations
• Optimal hypothesis testing for high dimensional covariance matrices
• Large sample test criteria on Toeplitz covariance structure

zbMATH Keywords

high-dimensional dataToeplitz matrixU-statisticcovariance matrixminimax hypothesis testingoptimal separation ratessharp asymptotic rates

Mathematics Subject Classification ID

Nonparametric hypothesis testing (62G10) Asymptotic properties of nonparametric inference (62G20) Multivariate distribution of statistics (62H10) Hypothesis testing in multivariate analysis (62H15)

Cites Work

• A partial overview of the theory of statistics with functional data
• Variable selection in infinite-dimensional problems
• Optimal hypothesis testing for high dimensional covariance matrices
• A new test for sphericity of the covariance matrix for high dimensional data
• Some hypothesis tests for the covariance matrix when the dimension is large compared to the sample size
• Nonparametric goodness-of-fit testing under Gaussian models
• Asymptotic theory for maximum deviations of sample covariance matrix estimates
• New weighted portmanteau statistics for time series goodness of fit testing
• Tests for high-dimensional covariance matrices
• Limiting laws of coherence of random matrices with applications to testing covariance structure and construction of compressed sensing matrices
• Corrections to LRT on large-dimensional covariance matrix by RMT
• Contributions in infinite-dimensional statistics and related topics. Selected papers from the 3rd international workshop on functional and operatorial statistics (IWFOS'2014), Stresa, Italy, June 19--21, 2014
• Asymptotically minimax hypothesis testing for nonparametric alternatives. I
• SPECTRAL DENSITY ESTIMATION VIA NONLINEAR WAVELET METHODS FOR STATIONARY NON-GAUSSIAN TIME SERIES
• Asymptotic equivalence of spectral density estimation and Gaussian white noise
• A minimax test for hypotheses on a spectral density
• Asymptotically efficient estimation of analytic functions in Gaussian noise
• Adaptive estimation of the spectral density of a weakly or strongly dependent Gaussian process
• Optimal rates of convergence for estimating Toeplitz covariance matrices
• Tests for covariance matrices in high dimension with less sample size
• Minimax goodness-of-fit testing in multivariate nonparametric regression
• An exact test about the covariance matrix
• Sharp minimax tests for large covariance matrices and adaptation
• Likelihood ratio tests for covariance matrices of high-dimensional normal distributions
• Nonparametric Estimation of Smooth Spectral Densities of Gaussian Stationary Sequences
• Stringing high-dimensional data for functional analysis
• Quadratic functional estimation in inverse problems
• Adaptive estimation of the spectrum of a stationary Gaussian sequence
• On testing for an identity covariance matrix when the dimensionality equals or exceeds the sample size
• Test for bandedness of high-dimensional covariance matrices and bandwidth estimation
• A unified treatment for non-asymptotic and asymptotic approaches to minimax signal detection

Cited In (6)

• Sharp minimax tests for large covariance matrices and adaptation
• Fast nonasymptotic testing and support recovery for large sparse Toeplitz covariance matrices
• Adaptive test for large covariance matrices with missing observations
• Large sample test criteria on Toeplitz covariance structure
• An introduction to recent advances in high/infinite dimensional statistics
• Median-of-means approach for repeated measures data

Uses Software

• WeightedPortTest

This page was built for publication: Sharp minimax tests for large Toeplitz covariance matrices with repeated observations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q268748)