ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1107.0145
58
35

A CLT for Information-theoretic statistics of Non-centered Gram random matrices

1 July 2011
W. Hachem
M. Kharouf
J. Najim
J. W. Silverstein
ArXivPDFHTML
Abstract

In this article, we study the fluctuations of the random variable: {\mathcal I}_n(\rho) = \frac 1N \log\det(\Sigma_n \Sigma_n^* + \rho I_N),\quad (\rho>0) where Σn=n−1/2Dn1/2XnD~n1/2+An\Sigma_n= n^{-1/2} D_n^{1/2} X_n\tilde D_n^{1/2} +A_nΣn​=n−1/2Dn1/2​Xn​D~n1/2​+An​, as the dimensions of the matrices go to infinity at the same pace. Matrices XnX_nXn​ and AnA_nAn​ are respectively random and deterministic N×nN\times nN×n matrices; matrices DnD_nDn​ and D~n\tilde D_nD~n​ are deterministic and diagonal, with respective dimensions N×NN\times NN×N and n×nn\times nn×n; matrix Xn=(Xij)X_n=(X_{ij})Xn​=(Xij​) has centered, independent and identically distributed entries with unit variance, either real or complex. We prove that when centered and properly rescaled, the random variable In(ρ){\mathcal I}_n(\rho)In​(ρ) satisfies a Central Limit Theorem and has a Gaussian limit. The variance of In(ρ){\mathcal I}_n(\rho)In​(ρ) depends on the moment \EXij2\E X_{ij}^2\EXij2​ of the variables XijX_{ij}Xij​ and also on its fourth cumulant κ=\E∣Xij∣4−2−∣\EXij2∣2\kappa= \E|X_{ij}|^4 - 2 - |\E X_{ij}^2|^2κ=\E∣Xij​∣4−2−∣\EXij2​∣2. The main motivation comes from the field of wireless communications, where In(ρ){\mathcal I}_n(\rho)In​(ρ) represents the mutual information of a multiple antenna radio channel. This article closely follows the companion article "A CLT for Information-theoretic statistics of Gram random matrices with a given variance profile", {\em Ann. Appl. Probab. (2008)} by Hachem et al., however the study of the fluctuations associated to non-centered large random matrices raises specific issues, which are addressed here.

View on arXiv
Comments on this paper