Asymptotic power of sphericity tests for high-dimensional data

Abstract

This paper studies the asymptotic power of tests of sphericity against perturbations in a single unknown direction as both the dimensionality of the data and the number of observations go to infinity. We establish the convergence, under the null hypothesis and contiguous alternatives, of the log ratio of the joint densities of the sample covariance eigenvalues to a Gaussian process indexed by the norm of the perturbation. When the perturbation norm is larger than the phase transition threshold studied in Baik, Ben Arous and Peche [Ann. Probab. 33 (2005) 1643-1697] the limiting process is degenerate, and discrimination between the null and the alternative is asymptotically certain. When the norm is below the threshold, the limiting process is nondegenerate, and the joint eigenvalue densities under the null and alternative hypotheses are mutually contiguous. Using the asymptotic theory of statistical experiments, we obtain asymptotic power envelopes and derive the asymptotic power for various sphericity tests in the contiguity region. In particular, we show that the asymptotic power of the Tracy-Widom-type tests is trivial (i.e., equals the asymptotic size), whereas that of the eigenvalue-based likelihood ratio test is strictly larger than the size, and close to the power envelope.

Published in at http://dx.doi.org/10.1214/13-AOS1100 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (3)

• Finite sample approximation results for principal component analysis: a matrix perturbation approach
• Semiparametrically efficient rank-based inference for shape I. optimal rank-based tests for sphericity
• The rank 1 real Wishart spiked model

Cited by in corpus (30)

• Optimality and Sub-optimality of PCA I: Spiked Random Matrix Models
• Complex energy landscapes in spiked-tensor and simple glassy models: ruggedness, arrangements of local minima and phase transitions
• On high-dimensional sign tests
• A necessary and sufficient condition for edge universality at the largest singular values of covariance matrices
• Optimality and Sub-optimality of PCA for Spiked Random Matrices and Synchronization
• Fundamental limits of detection in the spiked Wigner model
• Optimal hypothesis testing for stochastic block models with growing degrees
• Sharp detection in PCA under correlations: all eigenvalues matter
• Edge universality of separable covariance matrices
• Subexponential-Time Algorithms for Sparse PCA
• Asymptotic Linear Spectral Statistics for Spiked Hermitian Random Matrix Models
• Phase transition in random tensors with multiple independent spikes
• Estimation of functionals of sparse covariance matrices
• Free energy fluctuations of the $2$-spin spherical SK model at critical temperature
• Detecting Rare and Weak Spikes in Large Covariance Matrices
• Adapting to Unknown Noise Distribution in Matrix Denoising
• $L^2$ Asymptotics for High-Dimensional Data
• Optimal Covariance Estimation for Condition Number Loss in the Spiked Model
• Weak Detection in the Spiked Wigner Model with General Rank
• Weak detection in the spiked Wigner model
• Three Problems Related to the Eigenvalues of Complex Non-central Wishart Matrices with a Rank-1 Mean
• Detection of Signal in the Spiked Rectangular Models
• The Spectral Norm of Random Inner-Product Kernel Matrices
• Asymptotic joint distribution of extreme eigenvalues and trace of large sample covariance matrix in a generalized spiked population model
• Long Random Matrices and Tensor Unfolding
• Joint density of eigenvalues in spiked multivariate models
• High Dimensional Rank Tests for Sphericity
• Massive MIMO as a Big Data System: Random Matrix Models and Testbed
• High-dimensional tests for spherical location and spiked covariance
• Precise Error Rates for Computationally Efficient Testing