Large deviations of spread measures for Gaussian matrices
Abstract
For a large Gaussian matrix, we compute the joint statistics, including large deviation tails, of generalized and total variance - the scaled log-determinant and trace of the corresponding covariance matrix. Using a Coulomb gas technique, we find that the Laplace transform of their joint distribution decays for large (with fixed) as , where is the Dyson index of the ensemble and is a -independent large deviation function, which we compute exactly for any . The corresponding large deviation functions in real space are worked out and checked with extensive numerical simulations. The results are complemented with a finite treatment based on the Laguerre-Selberg integral. The statistics of atypically small log-determinants is shown to be driven by the split-off of the smallest eigenvalue, leading to an abrupt change in the large deviation speed.
Cite
@article{arxiv.1403.4494,
title = {Large deviations of spread measures for Gaussian matrices},
author = {Fabio Deelan Cunden and Pierpaolo Vivo},
journal= {arXiv preprint arXiv:1403.4494},
year = {2016}
}
Comments
20 pages, 3 figures. v4: final version