Aggregation by exponential weighting and sharp oracle inequalities

被引：50

作者：

Dalalyan, Arnak S. ^{[1
]}

Tsybakov, Alexandre B. ^{[1
]}

机构：

[1] Univ Paris 06, 4 Pl Jussieu, F-75252 Paris 05, France

来源：

LEARNING THEORY, PROCEEDINGS | 2007年 / 4539卷

关键词：

D O I：

10.1007/978-3-540-72927-3_9

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In the present paper, we study the problem of aggregation under the squared loss in the model of regression with deterministic design. We obtain sharp oracle inequalities for convex aggregates defined via exponential weights, under general assumptions on the distribution of errors and on the functions to aggregate. We show how these results can be applied to derive a sparsity oracle inequality.

引用

页码：97 / +

页数：5

共 30 条

[1] A randomized online learning algorithm for better variance control [J].

Audibert, Jean-Yves .

LEARNING THEORY, PROCEEDINGS, 2006, 4005 :392-407

[2]

AUDIBERT JY, 2004, THESIS U PARIS

[3]

BUNEA F, 2005, SEQUENTIAL PROCEDURE

[4]

BUNEA F, 2006, UNPUB SPARSITY ORACL

[5]

BUNEA F, 2007, IN PRESS ANN STAT

[6] Aggregation and sparsity via l1 penalized least squares [J].

Bunea, Florentina ;

Tsybakov, Alexandre B. ;

Wegkamp, Marten H. .

LEARNING THEORY, PROCEEDINGS, 2006, 4005 :379-391

[7]

CANDES E, 2007, IN PRESS ANN STAT

[8]

Catoni O., 2004, LECT NOTES MATH

[9]

CATONI O, U AGGREGATION RULES

[10] On the generalization ability of on-line learning algorithms [J].

Cesa-Bianchi, N ;

Conconi, A ;

Gentile, C .

IEEE TRANSACTIONS ON INFORMATION THEORY, 2004, 50 (09) :2050-2057

← 1 2 3 →