Local Asymptotic Normality of the Spectrum of High-Dimensional Spiked F-Ratios

We consider two types of spiked multivariate F distributions: a scaled distribution with the scale matrix equal to a rank-k perturbation of the identity, and a distribution with trivial scale, but rank-k non-centrality. The eigenvalues of the rank-k matrix (spikes) parameterize the joint distribution of the eigenvalues of the corresponding F matrix. We show that, for the spikes located above a phase transition threshold, the asymptotic behavior of the log ratio of the joint density of the eigenvalues of the F matrix to their joint density under a local deviation from these values depends only on the k of the largest eigenvalues {code} Furthermore, we show that {code} are asymptotically jointly normal, and the statistical experiment of observing all the eigenvalues of the F matrix converges in the Le Cam sense to a Gaussian shift experiment that depends on the asymptotic means and variances of {code}. In particular, the best statistical inference about sufficiently large spikes in the local asymptotic regime is based on the k of the largest eigenvalues only.

[1]  J. W. Silverstein,et al.  No eigenvalues outside the support of the limiting spectral distribution of large-dimensional sample covariance matrices , 1998 .

[2]  Z. Bai,et al.  CLT for linear spectral statistics of large dimensional sample covariance matrices with dependent data , 2017, Statistical Papers.

[3]  Alex Bloemendal,et al.  Limits of spiked random matrices II , 2011, 1109.3704.

[4]  A. James.  Distributions of Matrix Variates and Latent Roots Derived from Normal Samples , 1964 .

[5]  A. V. D. Vaart,et al.  Asymptotic Statistics: U -Statistics , 1998 .

[6]  Raj Rao Nadakuditi,et al.  Fundamental Limit of Sample Generalized Eigenvalue Based Detection of Signals in Noise Using Relatively Few Signal-Bearing and Noise-Only Samples , 2009, IEEE Journal of Selected Topics in Signal Processing.

[7]  R. Muirhead.  Aspects of Multivariate Statistical Theory , 1982, Wiley Series in Probability and Statistics.

[8]  Muni S. Srivastava,et al.  Estimation of the eigenvalues of σ 1 σ 2 -1 , 1992 .

[9]  Wang Zhou,et al.  Universality for the largest eigenvalue of sample covariance matrices with general population , 2013, 1304.5690.

[10]  Jianfeng Yao,et al.  Extreme eigenvalues of large-dimensional spiked Fisher matrices with application , 2015, 1504.05087.

[11]  Matthew R. McKay,et al.  Asymptotic Linear Spectral Statistics for Spiked Hermitian Random Matrices , 2014, 1402.6419.

[12]  S. Péché,et al.  Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices , 2004, math/0403022.

[13]  A. Onatski.  Detection of weak signals in high-dimensional complex-valued data , 2012, 1207.7098.

[14]  J. W. Silverstein.  Strong convergence of the empirical distribution of eigenvalues of large dimensional random matrices , 1995 .

[15]  J. W. Silverstein,et al.  On the empirical distribution of eigenvalues of a class of large dimensional random matrices , 1995 .

[16]  Marcelo J. Moreira,et al.  Signal detection in high dimension , 2012 .

[17]  I. Johnstone.  On the distribution of the largest eigenvalue in principal components analysis , 2001 .

[18]  Charles R. Johnson,et al.  Matrix analysis , 1985, Statistical Inference for Engineers and Data Scientists.

[19]  J. W. Silverstein,et al.  Analysis of the limiting spectral distribution of large dimensional random matrices , 1995 .

[20]  A. Guionnet,et al.  Fluctuations of the Extreme Eigenvalues of Finite Rank Deformations of Random Matrices , 2010, 1009.0145.

[21]  F. Olver.  Asymptotics and Special Functions , 1974 .

[22]  S. P'ech'e,et al.  The largest eigenvalues of sample covariance matrices for a spiked population: Diagonal case , 2008, 0812.2320.

[23]  W. G. Hill,et al.  Modification of Estimates of Parameters in the Construction of Genetic Selection Indices ('Bending') , 1981 .

[24]  Alexei Onatski,et al.  Signal detection in high dimension: The multispiked case , 2012, 1210.5663.

[25]  Marcelo J. Moreira,et al.  Asymptotic power of sphericity tests for high-dimensional data , 2013, 1306.4867.

[26]  Z. Bai,et al.  Central limit theorems for eigenvalues in a spiked population model , 2008, 0806.2503.

[27]  D. Paul,et al.  Asymptotics of the leading sample eigenvalues for a spiked covariance model , 2004 .

[28]  P. Hall,et al.  Martingale Limit Theory and its Application. , 1984 .

[29]  Alexei Onatski,et al.  Asymptotics of the principal components estimator of large factor models with weakly influential factors , 2012 .

[30]  K. Wachter.  The Limiting Empirical Measure of Multiple Discriminant Ratios , 1980 .

[31]  D. McLeish.  Dependent Central Limit Theorems and Invariance Principles , 1974 .

[32]  Mark Kritzman,et al.  Skulls, Financial Turbulence, and Risk Management , 2010 .

[33]  Yasunori Fujikoshi,et al.  Estimation of the eigenvalues of noncentrality parameter in matrix variate noncentral beta distribution , 2004 .

[34]  P. Hall,et al.  Martingale Limit Theory and Its Application , 1980 .

[35]  Alex Bloemendal,et al.  Limits of spiked random matrices I , 2010, Probability Theory and Related Fields.

[36]  M. Y. Mo.  Rank 1 real Wishart spiked model , 2011, 1101.5144.

[37]  A. Onatski.  Asymptotics of the principal components estimator of large factor models with weak factors and i.i.d. Gaussian noise , 2018 .

[38]  C. Donati-Martin,et al.  The largest eigenvalues of finite rank deformation of large Wigner matrices: Convergence and nonuniversality of the fluctuations. , 2007, 0706.0136.

[39]  Jiang Hu,et al.  Canonical correlation coefficients of high-dimensional Gaussian vectors: Finite rank case , 2014, The Annals of Statistics.

[40]  Tseng C. Chang.  On an Asymptotic Representation of the Distribution of the Characteristic Roots of $S_1S_2^{-1}$ , 1970 .

[41]  Matthew R. McKay,et al.  Hypergeometric functions of matrix arguments and linear statistics of multi-spiked Hermitian matrix models , 2014, J. Multivar. Anal..

[42]  J. W. Silverstein.  The Limiting Eigenvalue Distribution of a Multivariate F Matrix , 1985 .

[43]  Tosio Kato.  Perturbation theory for linear operators , 1966 .

[44]  William J. Glynn.  Asymptotic Representations of the Densities of Canonical Correlations and Latent Roots in MANOVA When the Population Parameters Have Arbitrary Multiplicity , 1980 .

[45]  Prathapasinghe Dharmawansa,et al.  Joint density of eigenvalues in spiked multivariate models , 2014, Stat.

[46]  L. C. Hsu.  A theorem on the asymptotic behavior of a multiple integral , 1948 .

[47]  J. Magnus,et al.  Matrix Differential Calculus with Applications in Statistics and Econometrics , 1991 .

[48]  Z. Bai,et al.  Limiting behavior of the eigenvalues of a multivariate F matrix , 1983 .