60
31

Extreme eigenvalues of large-dimensional spiked Fisher matrices with application

Abstract

Consider two pp-variate populations, not necessarily Gaussian, with covariance matrices Σ1\Sigma_1 and Σ2\Sigma_2, respectively, and let S1S_1 and S2S_2 be the sample covariances matrices from samples of the populations with degrees of freedom TT and nn, respectively. When the difference Δ\Delta between Σ1\Sigma_1 and Σ2\Sigma_2 is of small rank compared to p,Tp,T and nn, the Fisher matrix F=S21S1F=S_2^{-1}S_1 is called a {\em spiked Fisher matrix}. When p,Tp,T and nn grow to infinity proportionally, we establish a phase transition for the extreme eigenvalues of FF: when the eigenvalues of Δ\Delta ({\em spikes}) are above (or under) a critical value, the associated extreme eigenvalues of the Fisher matrix will converge to some point outside the support of the global limit (LSD) of other eigenvalues; otherwise, they will converge to the edge points of the LSD. Furthermore, we derive central limit theorems for these extreme eigenvalues of the spiked Fisher matrix. The limiting distributions are found to be Gaussian if and only if the corresponding population spike eigenvalues in Δ\Delta are {\em simple}. Numerical examples are provided to demonstrate the finite sample performance of the results. In addition to classical applications of a Fisher matrix in high-dimensional data analysis, we propose a new method for the detection of signals allowing an arbitrary covariance structure of the noise. Simulation experiments are conducted to illustrate the performance of this detector.

View on arXiv
Comments on this paper