Nous donnons des constantes dans l'inégalité de concentration à gauche de Talagrand pour les processus empiriques indexés par des classes de fonctions, en partant de la méthode de Herbst. Le point nouveau est que la constante du facteur variance est exacte, ce qui répond à une conjecture de Massart.
We give new constants in Talagrand's left concentration inequality for maxima of empirical processes. Our approach is based on the Herbst method. The improvement we get concerns the constant in the variance factor, which is the one conjectured by Massart.
Accepté le :
Publié le :
Thierry Klein 1
@article{CRMATH_2002__334_6_501_0, author = {Thierry Klein}, title = {Une in\'egalit\'e de concentration \`a gauche pour les processus empiriques}, journal = {Comptes Rendus. Math\'ematique}, pages = {501--504}, publisher = {Elsevier}, volume = {334}, number = {6}, year = {2002}, doi = {10.1016/S1631-073X(02)02303-8}, language = {fr}, }
Thierry Klein. Une inégalité de concentration à gauche pour les processus empiriques. Comptes Rendus. Mathématique, Volume 334 (2002) no. 6, pp. 501-504. doi : 10.1016/S1631-073X(02)02303-8. https://comptes-rendus.academie-sciences.fr/mathematique/articles/10.1016/S1631-073X(02)02303-8/
[1] On Talagrand's deviation inequalities for product measures, ESAIM Probability and Statistics, Volume 1 (1996), pp. 63-87
[2] About the constant in Talagrand's concentration inequalities for empirical processes, Ann. Probab., Volume 28 (2000), pp. 863-884
[3] Théorie asymptotique des processus aléatoires faiblement dépendants (J.M. Ghidaglia; X. Guyon, eds.), Math. Appl., 31, Springer, 2000
[4] E. Rio, Une inégalité de Bernstein pour les maxima de processus empiriques, Preprint 57 de l'Université de Versailles–Saint-Quentin, 2001
[5] New concentration inequalities in product spaces, Invent. Math., Volume 126 (1996), pp. 503-563
- Concentration inequalities for suprema of unbounded empirical processes, Annales Henri Lebesgue, Volume 4 (2021), p. 831 | DOI:10.5802/ahl.90
- About the rate function in concentration inequalities for suprema of bounded empirical processes, Stochastic Processes and their Applications, Volume 129 (2019) no. 10, p. 3967 | DOI:10.1016/j.spa.2018.11.010
- On optimality of empirical risk minimization in linear aggregation, Bernoulli, Volume 24 (2018) no. 3 | DOI:10.3150/17-bej925
- On Concentration for (Regularized) Empirical Risk Minimization, Sankhya A, Volume 79 (2017) no. 2, p. 159 | DOI:10.1007/s13171-017-0111-9
- Global Rademacher Complexity Bounds: From Slow to Fast Convergence Rates, Neural Processing Letters, Volume 43 (2016) no. 2, p. 567 | DOI:10.1007/s11063-015-9429-2
- , 2015 International Joint Conference on Neural Networks (IJCNN) (2015), p. 1 | DOI:10.1109/ijcnn.2015.7280414
- Optimal upper and lower bounds for the true and empirical excess risks in heteroscedastic least-squares regression, Electronic Journal of Statistics, Volume 6 (2012) no. none | DOI:10.1214/12-ejs679
- Uniform in bandwidth exact rates for a class of kernel estimators, Annals of the Institute of Statistical Mathematics, Volume 63 (2011) no. 6, p. 1077 | DOI:10.1007/s10463-010-0286-5
- On the Optimality of Sample-Based Estimates of the Expectation of the Empirical Minimizer, ESAIM: Probability and Statistics, Volume 14 (2010), p. 315 | DOI:10.1051/ps:2008036
- A limited in bandwidth uniformity for the functional limit law of the increments of the empirical process, Electronic Journal of Statistics, Volume 2 (2008) no. none | DOI:10.1214/08-ejs193
- Fast rates for support vector machines using Gaussian kernels, The Annals of Statistics, Volume 35 (2007) no. 2 | DOI:10.1214/009053606000001226
- Empirical minimization, Probability Theory and Related Fields, Volume 135 (2006) no. 3, p. 311 | DOI:10.1007/s00440-005-0462-3
- Local Rademacher complexities and oracle inequalities in risk minimization, The Annals of Statistics, Volume 34 (2006) no. 6 | DOI:10.1214/009053606000001019
- Concentration around the mean for maxima of empirical processes, The Annals of Probability, Volume 33 (2005) no. 3 | DOI:10.1214/009117905000000044
- Concentration Inequalities, Advanced Lectures on Machine Learning, Volume 3176 (2004), p. 208 | DOI:10.1007/978-3-540-28650-9_9
- Local Complexities for Empirical Risk Minimization, Learning Theory, Volume 3120 (2004), p. 270 | DOI:10.1007/978-3-540-27819-1_19
- Ratio Limit Theorems for Empirical Processes, Stochastic Inequalities and Applications (2003), p. 249 | DOI:10.1007/978-3-0348-8069-5_15
- Symmetrization approach to concentration inequalities for empirical processes, The Annals of Probability, Volume 31 (2003) no. 4 | DOI:10.1214/aop/1068646378
Cité par 18 documents. Sources : Crossref
Commentaires - Politique