On the asymptotic behaviour of the eigenvalue distribution of block correlation matrices of high-dimensional time series

TL;DR

This paper investigates the asymptotic eigenvalue distribution of block correlation matrices derived from high-dimensional time series, showing convergence to the Marcenko-Pastur law, which aids in testing uncorrelatedness.

Contribution

It provides a theoretical analysis of the eigenvalue distribution of block correlation matrices in high-dimensional settings, extending understanding of their asymptotic behavior.

Findings

Eigenvalue distribution converges to Marcenko-Pastur law.

Asymptotic regime where M, L, N grow with fixed ratio.

Results useful for testing independence among many time series.

Abstract

We consider linear spectral statistics built from the block-normalized correlation matrix of a set of $M$ mutually independent scalar time series. This matrix is composed of $M \times M$ blocks that contain the sample cross correlation between pairs of time series. In particular, each block has size $L \times L$ and contains the sample cross-correlation measured at $L$ consecutive time lags between each pair of time series. Let $N$ denote the total number of consecutively observed windows that are used to estimate these correlation matrices. We analyze the asymptotic regime where $M,L,N \rightarrow +\infty$ while $ML/N \rightarrow c_\star$, $0<c_\star<\infty$. We study the behavior of linear statistics of the eigenvalues of this block correlation matrix under these asymptotic conditions and show that the empirical eigenvalue distribution converges to a Marcenko-Pastur distribution. Our results are potentially useful in order to address the problem of testing whether a large number of time series are uncorrelated or not.