On asymptotics of eigenvectors of large sample covariance matrix

Random matrices (algebraic aspects) (15B52) Central limit and other weak theorems (60F05) Functional limit theorems; invariance principles (60F17) Asymptotic distribution theory in statistics (62E20)

Abstract: Let {

X_{i j}

},

i, j = ...,

be a double array of i.i.d. complex random variables with

E X_{11} = 0, E | X_{11} |^{2} = 1

and

E | X_{11} |^{4} < i n f t y

, and let

A_{n} = f r a c 1 N T_{n}^{1 / 2} X_{n} X_{n}^{*} T_{n}^{1 / 2}

, where

T_{n}^{1 / 2}

is the square root of a nonnegative definite matrix

T_{n}

and

X_{n}

is the

n i m e s N

matrix of the upper-left corner of the double array. The matrix

A_{n}

can be considered as a sample covariance matrix of an i.i.d. sample from a population with mean zero and covariance matrix

T_{n}

, or as a multivariate

F

matrix if

T_{n}

is the inverse of another sample covariance matrix. To investigate the limiting behavior of the eigenvectors of

A_{n}

, a new form of empirical spectral distribution is defined with weights defined by eigenvectors and it is then shown that this has the same limiting spectral distribution as the empirical spectral distribution defined by equal weights. Moreover, if {

X_{i j}

} and

T_{n}

are either real or complex and some additional moment assumptions are made then linear spectral statistics defined by the eigenvectors of

A_{n}

are proved to have Gaussian limits, which suggests that the eigenvector matrix of

A_{n}

is nearly Haar distributed when

T_{n}

is a multiple of the identity matrix, an easy consequence for a Wishart matrix.

Review text

Let \(X_n=(X_{ij})\) be an \(n\times N\) matrix of independent, identically distributed complex random variables and \(T_n\) be an \(n\times n\) nonnegative definite Hermitian matrix with a square root \(T_n^{1/2}\), and \(A_n=\displaystyle\frac{1}{N}T_n^{1/2}X_nX_n^*T_n^{1/2}\), in which both the dimension \(n\) and the sample size \(N\) are large. The authors show the central limit theorem for those analytic functions over the support of the limiting spectral distribution of \(A_n\) under the condition that \[ EX_{11}=0,\quad E|X_{11}|^2=1, \quad E|X_{11}|^4<\infty. \] Moreover, if \(\{X_{ij}\}\) and \(T_n\) are either real or complex and some additional moment assumptions are made then the linear spectral statistic defined by the eigenvectors of \(A_n\) are proved to have Gaussian limits, which suggests that the eigenvector matrix of \(A_n\) is nearly Haar distributed when \(T_n\) is a multiple of the identity matrix, an easy consequence for a Wishart matrix.

Recommendations

Cites work

Cited in

(71)

This page was built for publication: On asymptotics of eigenvectors of large sample covariance matrix

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2373573)