Local Rademacher complexities
作者： Peter L. Bartlett , Olivier Bousquet , Shahar Mendelson
DOI: 10.1214/009053605000000282
关键词:
摘要: We propose new bounds on the error of learning algorithms in terms a data-dependent notion complexity. The estimates we establish give optimal rates and are based local empirical version Rademacher averages, sense that averages computed from data, subset functions with small error. present some applications to classification prediction convex function classes, kernel classes particular.
unirioja.es
berkeley.edu
core.ac.uk
mpg.de参考文章(34)
László Györfi, Michael Kohler, Adam Krzyżak, Harro Walk, A distribution-free theory of nonparametric regression Published in <b>2002</b> in New York NY) by Springer. ,(2002) , 10.1007/B97848
Pascal Massart, About the constants in Talagrand's concentration inequalities for empirical processes Annals of Probability. ,vol. 28, pp. 863- 884 ,(2000) , 10.1214/AOP/1019160263
Richard M. Dudley, Uniform Central Limit Theorems ,(1999)
Olivier Bousquet, Vladimir Koltchinskii, Dmitriy Panchenko, Some Local Measures of Complexity of Convex Hulls and Generalization Bounds conference on learning theory. pp. 59- 73 ,(2002) , 10.1007/3-540-45435-7_5
Olivier Bousquet, Concentration Inequalities for Sub-Additive Functions Using the Entropy Method Stochastic Inequalities and Applications. ,vol. 56, pp. 213- 247 ,(2003) , 10.1007/978-3-0348-8069-5_14
Michel Ledoux, Michel Talagrand, Probability in Banach Spaces: Isoperimetry and Processes ,(1991)
László Györfi, Luc Devroye, Gábor Lugosi, A Probabilistic Theory of Pattern Recognition ,(1996)
Shahar Mendelson, Geometric Parameters of Kernel Machines conference on learning theory. pp. 29- 43 ,(2002) , 10.1007/3-540-45435-7_3
Luc Devroye, Gábor Lugosi, Combinatorial Methods in Density Estimation ,(2011)
Vladimir Koltchinskii, Dmitriy Panchenko, Rademacher Processes and Bounding the Risk of Function Learning arXiv: Probability. pp. 443- 457 ,(2000) , 10.1007/978-1-4612-1358-1_29