English

Central limit theorem for eigenvalue statistics of sample covariance matrix with random population

Probability 2023-02-27 v2

Abstract

Consider the sample covariance matrix Σ1/2XXTΣ1/2\Sigma^{1/2}XX^T\Sigma^{1/2} where XX is an M×NM\times N random matrix with independent entries and Σ\Sigma is an M×MM\times M diagonal matrix. It is known that if Σ\Sigma is deterministic, then the fluctuation of if(λi)\sum_if(\lambda_i) converges in distribution to a Gaussian distribution. Here {λi}\{\lambda_i\} are eigenvalues of Σ1/2XXTΣ1/2\Sigma^{1/2}XX^T\Sigma^{1/2} and ff is a good enough test function. In this paper we consider the case that Σ\Sigma is random and show that the fluctuation of 1Nif(λi)\frac{1}{\sqrt N}\sum_if(\lambda_i) converges in distribution to a Gaussian distribution. This phenomenon implies that the randomness of Σ\Sigma decreases the correlation among {λi}\{\lambda_i\}.

Keywords

Cite

@article{arxiv.2211.05546,
  title  = {Central limit theorem for eigenvalue statistics of sample covariance matrix with random population},
  author = {Ji Oon Lee and Yiting Li},
  journal= {arXiv preprint arXiv:2211.05546},
  year   = {2023}
}

Comments

Errors corrected. More details included. Main theorem strengthened

R2 v1 2026-06-28T05:35:48.205Z