Adaptive test for large covariance matrices with missing observations

MaRDI QIDQ5278118zbMATH OpenFDO

Authors Cristina Butucea, Rania Zgheib

Publication date 13 July 2017

Full work available at URL https://arxiv.org/abs/1602.04310, http://alea.impa.br/articles/v14/14-28.pdf

zbMATH Keywords

adaptive testcovariance matricesgoodness-of-fit testsminimax separation ratemissing observationToeplitz matrices

Mathematics Subject Classification ID

Nonparametric hypothesis testing (62G10) Hypothesis testing in multivariate analysis (62H15)

Abstract: We observe $n$ independent $p-$dimensional Gaussian vectors with missing coordinates, that is each value (which is assumed standardized) is observed with probability $a>0$. We investigate the problem of minimax nonparametric testing that the high-dimensional covariance matrix $Sigma$ of the underlying Gaussian distribution is the identity matrix, using these partially observed vectors. Here, $n$ and $p$ tend to infinity and $a>0$ tends to 0, asymptotically. We assume that $Sigma$ belongs to a Sobolev-type ellipsoid with parameter $alpha>0$. When $alpha$ is known, we give asymptotically minimax consistent test procedure and find the minimax separation rates $ildevarph{i}_{n,p}=(a^{2}nsqrtp{)}^{-frac2alpha4alpha+1}$, under some additional constraints on $n,,p$ and $a$. We show that, in the particular case of Toeplitz covariance matrices,the minimax separation rates are faster, $ildeph{i}_{n,p}=(a^{2}np{)}^{-frac2alpha4alpha+1}$. We note how the "missingness" parameter $a$ deteriorates the rates with respect to the case of fully observed vectors ($a=1$). We also propose adaptive test procedures, that is free of the parameter $alpha$ in some interval, and show that the loss of rate is $(lnln(a^{2}nsqrtp){)}^{alpha/(4alpha+1)}$ and $(lnln(a^{2}np){)}^{alpha/(4alpha+1)}$ for Toeplitz covariance matrices, respectively.

Recommendations

• Sharp minimax tests for large covariance matrices and adaptation
• Adaptive Tests for Bandedness of High-dimensional Covariance Matrices
• Covariance Matrix Estimation With Non Uniform and Data Dependent Missing Observations
• An adaptive-to-model test for parametric single-index models with missing responses
• Test for high-dimensional mean vector under missing observations
• Missing data in the k-population multivariate normal patterned mean and covariance matrix testing and estimation problem
• High-dimensional covariance matrix estimation with missing observations
• Simultaneous testing of the mean vector and the covariance matrix with-two-step monotone missing data
• A note on testing the covariance matrix for large dimension

Cites work

• Asymptotically minimax hypothesis testing for nonparametric alternatives. III
• Corrections to LRT on large-dimensional covariance matrix by RMT
• High-dimensional covariance matrix estimation with missing observations
• scientific article; zbMATH DE number 4084757 (Why is no real title available?)
• scientific article; zbMATH DE number 3723610 (Why is no real title available?)
• Introduction to nonparametric estimation
• Likelihood ratio tests for covariance matrices of high-dimensional normal distributions
• Nonparametric goodness-of-fit testing under Gaussian models
• Optimal hypothesis testing for high dimensional covariance matrices
• Sharp minimax tests for large covariance matrices and adaptation
• Sharp minimax tests for large Toeplitz covariance matrices with repeated observations
• Some hypothesis tests for the covariance matrix when the dimension is large compared to the sample size
• Spectral analysis of high-dimensional sample covariance matrices with missing observations
• Tests for high-dimensional covariance matrices

Cited in

(4)

• Fast nonasymptotic testing and support recovery for large sparse Toeplitz covariance matrices
• Sharp minimax tests for large Toeplitz covariance matrices with repeated observations
• Sharp minimax tests for large covariance matrices and adaptation
• Many-sample tests for the equality and the proportionality hypotheses between large covariance matrices

Describes a project that uses

Uses Software

• RegEM

This page was built for publication: Adaptive test for large covariance matrices with missing observations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5278118)