English

Extreme Eigenvalues of Large Dimensional Quaternion Sample Covariance Matrix

Probability 2013-12-18 v1

Abstract

In this paper, we shall investigate the almost sure limits of the largest and smallest eigenvalues of a quaternion sample covariance matrix. Suppose that Xn\mathbf X_n is a p×np\times n matrix whose elements are independent quaternion variables with mean zero, variance 1 and uniformly bounded fourth moments. Denote Sn=1nXnXn\mathbf S_n=\frac{1}{n}\mathbf X_n\mathbf X_n^*. In this paper, we shall show that smax(Sn)=sp(Sn)(1+y)2,a.s.s_{\max}\left(\mathbf S_n\right)=s_{p}\left(\mathbf S_n\right)\to\left(1+\sqrt y\right)^2, a.s. and smin(Sn)(1y)2,a.s.s_{\min}\left(\mathbf S_n\right)\to\left(1-\sqrt y\right)^2,a.s. as nn\to\infty, where y=limp/ny=\lim p/n, s1(Sn)sp(Sn)s_1\left(\mathbf S_n\right)\le\cdots\le s_{p}\left(\mathbf S_n\right) are the eigenvalues of Sn\mathbf{S}_n, smin(Sn)=spn+1(Sn)s_{\min}\left(\mathbf S_n\right)=s_{p-n+1}\left(\mathbf S_n\right) when p>np>n and smin(Sn)=s1(Sn)s_{\min}\left(\mathbf S_n\right)=s_1\left(\mathbf S_n\right) when pnp\le n. We also prove that the set of conditions are necessary for smax(Sn)(1+y)2,a.s.s_{\max}\left(\mathbf S_n\right)\to\left(1+\sqrt y\right)^2, a.s. when the entries of Xn\mathbf {X}_n are i. i. d.

Keywords

Cite

@article{arxiv.1312.4649,
  title  = {Extreme Eigenvalues of Large Dimensional Quaternion Sample Covariance Matrix},
  author = {Huiqin Li and Zhidong Bai},
  journal= {arXiv preprint arXiv:1312.4649},
  year   = {2013}
}

Comments

21 pages

R2 v1 2026-06-22T02:29:07.876Z