Spectral deconvolution of unitarily invariant matrix models

DOI10.1093/IMAIAI/IAAD037MaRDI QIDQ6343061zbMATH OpenOpenAlexFDO

Authors Pierre Tarrago

Publication date 16 June 2020

Full work available at URL https://doi.org/10.1093/imaiai/iaad037

Mathematics Subject Classification ID

Random matrices (algebraic aspects) (15B52) Exact distribution theory in statistics (62E15) Signal theory (characterization, reconstruction, filtering, etc.) (94A12)

Abstract: The present paper implements a complex analytic method to recover the spectrum of a matrix perturbed by either the addition or the multiplication of a random matrix noise, under the assumption that the distribution of the noise is unitarily invariant. This method, introduced by Arizmendi, Tarrago and Vargas in arXiv:1711.08871, is done in two steps : the first step consists in a fixed point method to compute the Stieltjes transform of the desired distribution in a certain domain, and the second step is a classical deconvolution by a Cauchy distribution, whose parameter depends on the intensity of the noise. We also provide explicit bounds for the mean squared error of the first step.

Recommendations

• Sparse covariance matrix estimation in high-dimensional deconvolution
• A spectral method for deconvolving a density
• On the spectrum of noisy blown-up matrices
• Empirical spectral distribution of a matrix under perturbation
• Freeness over the diagonal and outliers detection in deformed random matrices with a variance profile

Cited in

(2)

• Analytic continuation from limited noisy Matsubara data
• Matrix denoising: Bayes-optimal estimators via low-degree polynomials

This page was built for publication: Spectral deconvolution of unitarily invariant matrix models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6343061)