Abstract
Let $X=(X_{i,j})_{m\times n}, m\ge n$, be a complex Gaussian random matrix with mean zero and variance $\frac 1n$, let $S=X^*X$ be a sample covariance matrix. In this paper we are mainly interested in the limiting behavior of eigenvalues when $\frac mn\rightarrow \gamma\ge 1$ as $n\rightarrow\infty$. Under certain conditions on $k$, we prove the central limit theorem holds true for the $k$-th largest eigenvalues $\lambda_{(k)}$ as $k$ tends to infinity as $n\rightarrow\infty$. The proof is largely based on the Costin-Lebowitz-Soshnikov argument and the asymptotic estimates for the expectation and variance of the number of eigenvalues in an interval. The standard technique for the RH problem is used to compute the exact formula and asymptotic properties for the mean density of eigenvalues. As a by-product, we obtain a convergence speed of the mean density of eigenvalues to the Marchenko-Pastur distribution density under the condition $|\frac mn-\gamma|=O(\frac 1n)$.
Highlights
Introduction and main resultsLet X = (Xi,j) be a complex m×n, m ≥ n, random matrix, the entries of which are independent complexGaussian random variables with mean zero and variance, namely Re(Xi,j ), I m(Xi,j )form a family of independent real Gaussian random variables each with mean value 0 and variance Let SX∗X, can be viewed as a sample covariance matrix of m samples of n dimensional random vectors and it is of fundamental importance in multivariate statistical analysis.The complex sample covariance matrices was first studied by Goodman (5) and Khatri (9)
Under mainly certain conditions on k, we prove the central limit theorem holds true for the k-th largest eigenvalues λ(k) as k tends to infinity as n → ∞
Form a family of independent real Gaussian random variables each with mean value 0 and variance can be viewed as a sample covariance matrix of m samples of n dimensional random vectors and it is of fundamental importance in multivariate statistical analysis
Summary
A recent remarkable work is on a limiting distribution of the largest eigenvalue. (2) It is remarkable that with regard to the Plancherel measure on the set of partitions λ of n, the rows λ1, λ2, λ3, · · · of λ behaves, suitably scaled, like the 1st, 2nd, 3rd and so on eigenvalues of a random matrix from the GUE. Where Φ(x) is the standard normal distribution function This in turn follows from the Costin-Lebowitz-Soshnikov theorem (2; 13) as long as. C for simplicity, which may take different values in different places
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.