Comptes Rendus
Functional Analysis/Probability Theory
Lower estimates for the singular values of random matrices
[Minorations des valeurs singulières de matrices aléatoires]
Comptes Rendus. Mathématique, Volume 342 (2006) no. 4, pp. 247-252.

Soit Γ une matrice n×n, ayant pour coefficients des variables aléatoires indépendantes et identiquement distribuées (i.i.d.) vérifiant une décroissance sous-gaussienne des queues. Dans ce travail, nous obtenons des minorations de type polynomial des valeurs singulières de Γ, valables avec une probabilité proche de 1. Nous montrons aussi que si A est une matrice N×n avec N>n, dont les coefficients sont des variables aléatoires sous-gaussiennes i.i.d., alors l'espace E=ARn vérifie avec une grande probablilité les conditions du théorème de Kashin, c'est à dire les normes 2N et 1N sont équivalentes sur E. De plus la distance entre ces normes dépend polynomialement de δ=(Nn)/n.

Let Γ be an n×n matrix, whose entries are independent identically distributed (i.i.d.) random variables satisfying the subgaussian tail estimate. We obtain polynomial type lower estimates of the singular numbers of Γ, which hold with probability close to 1. We also show that if A is an N×n matrix with N>n, whose entries are i.i.d. subgaussian random variables, then with high probability the space E=ARn satisfies the conditions of Kashin's theorem, i.e. the 2N and 1N norms are equivalent on E. Moreover the distance between these norms polynomially depends on δ=(Nn)/n.

Reçu le :
Accepté le :
Publié le :
DOI : 10.1016/j.crma.2005.11.013
Mark Rudelson 1

1 Department of Mathematics, University of Missouri, Columbia, MO 65211, USA
@article{CRMATH_2006__342_4_247_0,
     author = {Mark Rudelson},
     title = {Lower estimates for the singular values of random matrices},
     journal = {Comptes Rendus. Math\'ematique},
     pages = {247--252},
     publisher = {Elsevier},
     volume = {342},
     number = {4},
     year = {2006},
     doi = {10.1016/j.crma.2005.11.013},
     language = {en},
}
TY  - JOUR
AU  - Mark Rudelson
TI  - Lower estimates for the singular values of random matrices
JO  - Comptes Rendus. Mathématique
PY  - 2006
SP  - 247
EP  - 252
VL  - 342
IS  - 4
PB  - Elsevier
DO  - 10.1016/j.crma.2005.11.013
LA  - en
ID  - CRMATH_2006__342_4_247_0
ER  - 
%0 Journal Article
%A Mark Rudelson
%T Lower estimates for the singular values of random matrices
%J Comptes Rendus. Mathématique
%D 2006
%P 247-252
%V 342
%N 4
%I Elsevier
%R 10.1016/j.crma.2005.11.013
%G en
%F CRMATH_2006__342_4_247_0
Mark Rudelson. Lower estimates for the singular values of random matrices. Comptes Rendus. Mathématique, Volume 342 (2006) no. 4, pp. 247-252. doi : 10.1016/j.crma.2005.11.013. https://comptes-rendus.academie-sciences.fr/mathematique/articles/10.1016/j.crma.2005.11.013/

[1] S. Artstein, O. Friedland, V. Milman, Large sections of L1n spanned by random sign-vectors are isomorphic to Euclidean with polynomial dependence, Preprint

[2] Z.D. Bai; Y.Q. Yin Limit of the smallest eigenvalue of a large-dimensional sample covariance matrix, Ann. Probab., Volume 21 (1993) no. 3, pp. 1275-1294

[3] K. Davidson; S.J. Szarek Local operator theory, random matrices and Banach spaces, Handbook of the Geometry of Banach Spaces, vol. I, North-Holland, Amsterdam, 2001, pp. 317-366

[4] A.Yu. Garnaev; E.D. Gluskin The widths of a Euclidean ball, Dokl. Akad. Nauk SSSR, Volume 277 (1984), pp. 1048-1052 (in Russian). English translation Soviet Math. Dokl., 30, 1984, pp. 200-204

[5] B. Kashin The widths of certain finite-dimensional sets and classes of smooth functions, Izv. Akad. Nauk SSSR Ser. Mat., Volume 41 (1977), pp. 334-351 (in Russian)

[6] A.E. Litvak; A. Pajor; M. Rudelson; N. Tomczak-Jaegermann Smallest singular value of random matrices and geometry of random polytopes, Adv. Math., Volume 195 (2005) no. 2, pp. 491-523

[7] A.E. Litvak; A. Pajor; M. Rudelson; N. Tomczak-Jaegermann; R. Vershynin Random Euclidean embeddings in spaces of bounded volume ratio, C. R. Acad. Sci. Paris, Ser. I, Volume 339 (2004), pp. 33-38

[8] A.E. Litvak, A. Pajor, M. Rudelson, N. Tomczak-Jaegermann, R. Vershynin, Euclidean embeddings in spaces of finite volume ratio via random matrices, J. Reine Angew. Math., in press

[9] M. Rudelson, Invertibility of random matrices: norm of the inverse, Preprint

Cité par Sources :

Commentaires - Politique


Ces articles pourraient vous intéresser

Reconstruction and subgaussian processes

Shahar Mendelson; Alain Pajor; Nicole Tomczak-Jaegermann

C. R. Math (2005)


Random Euclidean embeddings in spaces of bounded volume ratio

Alexander Litvak; Alain Pajor; Mark Rudelson; ...

C. R. Math (2004)


The least singular value of a random square matrix is O(n1/2)

Mark Rudelson; Roman Vershynin

C. R. Math (2008)