Nonlinear component analysis as a kernel eigenvalue problem

被引：6065

作者：

Scholkopf, B ^{[1
]}

Smola, A

Muller, KR

机构：

[1] Max Planck Inst Biol Cybernet, D-72076 Tubingen, Germany

[2] GMD First Forschungszentrum Informat Tech, D-12489 Berlin, Germany

来源：

NEURAL COMPUTATION | 1998年 / 10卷 / 05期

关键词：

D O I：

10.1162/089976698300017467

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

A new method for performing a nonlinear form of principal component analysis is proposed. By the use of integral operator kernel functions, one can efficiently compute principal components in high-dimensional feature spaces, related to input space by some nonlinear map-for instance, the space of all possible five-pixel products in 16 x 16 images. We give the derivation of the method and present experimental results on polynomial feature extraction for pattern recognition.

引用

页码：1299 / 1319

页数：21

共 18 条

[1]

AIZERMAN MA, 1965, AUTOMAT REM CONTR+, V25, P821

[2]

[Anonymous], 44 M PLANCK I BIOL K

[3]

Boser B. E., 1992, Proceedings of the Fifth Annual ACM Workshop on Computational Learning Theory, P144, DOI 10.1145/130385.130401

[4]

BREGLER C, 1994, ADV NEURAL INFORMATI, V6

[5]

Burges C., 1996, P 13 INT C MACH LEAR

[6] SUPPORT-VECTOR NETWORKS [J].

CORTES, C ;

VAPNIK, V .

MACHINE LEARNING, 1995, 20 (03) :273-297

[7]

COURANT R, 1953, METHODS MATH PHYSICS, V1

[8]

Diamantaras KI, 1996, Principal Component Neural Networks: Theory and Applications

[9]

Dunford N., 1963, LINEAR OPERATORS

[10] PRINCIPAL CURVES [J].

HASTIE, T ;

STUETZLE, W .

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1989, 84 (406) :502-516

← 1 2 →