Aggregation and sparsity via l1 penalized least squares

被引：38

作者：

Bunea, Florentina ^{[1
]}

Tsybakov, Alexandre B.

Wegkamp, Marten H.

机构：

[1] Florida State Univ, Dept Stat, Tallahassee, FL 32306 USA

[2] Univ Paris 06, Lab Probabilites & Modeles Aleatoires, F-75252 Paris 05, France

[3] Inst Informat Transmiss Problems, Moscow, Russia

来源：

LEARNING THEORY, PROCEEDINGS | 2006年 / 4005卷

关键词：

D O I：

10.1007/11776420_29

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper shows that near optimal rates of aggregation and adaptation to unknown sparsity can be simultaneously achieved via, penalized least squares in a nonparametric regression setting. The main tool is a novel oracle inequality on the sum between the empirical squared loss of the penalized least squares estimate and a term reflecting the sparsity of the unknown regression function.

引用

页码：379 / 391

页数：13

共 18 条

[1]

BARAUD Y, 2002, ESAIM-PROBAB STAT, V7, P127

[2] Minimum contrast estimators on sieves: exponential bounds and rates of convergence [J].

Birge, L ;

Massart, P .

BERNOULLI, 1998, 4 (03) :329-375

[3]

BIRGE L, 2002, PREPUBLICATION U PAR, V783

[4]

BUNEA F, 2004, CANADIAN J STAT, V22, P1

[5]

BUNEA F, 2005, AGGREGATION GAUSSIAN

[6]

CANDES EJ, 2005, DANTZIG SELECTOR STA

[7]

Catoni O., 2004, LECT NOTES MATH

[8] IDEAL SPATIAL ADAPTATION BY WAVELET SHRINKAGE [J].

DONOHO, DL ;

JOHNSTONE, IM .

BIOMETRIKA, 1994, 81 (03) :425-455

[9]

Gyorfi L., 2002, DISTRIBUTION FREE TH

[10]

Juditsky A, 2000, ANN STAT, V28, P681

← 1 2 →