Abstract
We consider a data matrix from a multivariate stationary process with a separable covariance function, where is a positive semi-definite matrix, Z a random matrix of uncorrelated standardized white noise, and a Toeplitz matrix. Under the assumption of long range dependence (LRD), we re-examine the consistency of two toeplitzifized estimators (unbiased) and (biased) for , which are known to be norm consistent with when the process is short range dependent (SRD). However in the LRD case, some simulations suggest that the norm consistency does not hold in general for both estimators. Instead, a weaker ratio consistency is established for the unbiased estimator , and a further weaker ratio LSD consistency is established for the biased estimator . The main result leads to a consistent whitening procedure on the original data matrix X, which is further applied to two real world questions, one is a signal detection problem, and the other is PCA on the space covariance to achieve a noise reduction and data compression.
Funding Statement
This work was supported by Department of Statistics and Actuarial Science of the University of Hong Kong.
Acknowledgments
We would like to thank Professor Romain Couillet in University of Grenoble-Alpes for fruitful discussions, and also thanks to the anonymous reviewers for their useful suggestions and comments.
Citation
Peng Tian. Jianfeng Yao. "Ratio-consistent estimation for long range dependent Toeplitz covariance with application to matrix data whitening." Electron. J. Statist. 16 (2) 5035 - 5079, 2022. https://doi.org/10.1214/22-EJS2060
Information