SVD algorithms: APEX-like versus subspace methods

被引：10

作者：

Weingessel, A ^{[1
]}

Hornik, K ^{[1
]}

机构：

[1] VIENNA TECH UNIV,INST STAT & WAHRSCHEINLICHKEITSTHEORIE,A-1040 VIENNA,AUSTRIA

来源：

NEURAL PROCESSING LETTERS | 1997年 / 5卷 / 03期

关键词：

APEX algorithm; principal component analysis; singular value decomposition; subspace algorithm;

D O I：

10.1023/A:1009642710601

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We compare several new SVD learning algorithms which are based on the subspace method in principal component analysis with the APEX-like algorithm proposed by Diamantaras. It is shown experimentally that the convergence of these algorithms is as fast as the convergence of the APEX-like algorithm.

引用

页码：177 / 184

页数：8

共 13 条

[1] LEARNING IN LINEAR NEURAL NETWORKS - A SURVEY [J].

BALDI, PF ;

HORNIK, K .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 1995, 6 (04) :837-858

[2] DYNAMIC-SYSTEMS THAT SORT LISTS, DIAGONALIZE MATRICES, AND SOLVE LINEAR-PROGRAMMING PROBLEMS [J].

BROCKETT, RW .

LINEAR ALGEBRA AND ITS APPLICATIONS, 1991, 146 :79-91

[3]

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

[4]

DIAMANTARAS KI, 1992, THESIS PRINCETON U

[5]

Golub GH, 1989, MATRIX COMPUTATIONS

[6]

Goodall I.H., 1990, BIDDLE, V1990, P861

[7]

KAY J, 1992, P INT JOINT C NEUR N, V4, P79

[8]

Oja E., 1989, International Journal of Neural Systems, V1, P61, DOI 10.1142/S0129065789000475

[9] A SIMPLIFIED NEURON MODEL AS A PRINCIPAL COMPONENT ANALYZER [J].

OJA, E .

JOURNAL OF MATHEMATICAL BIOLOGY, 1982, 15 (03) :267-273

[10] LYAPUNOV FUNCTIONS FOR CONVERGENCE OF PRINCIPAL COMPONENT ALGORITHMS [J].

PLUMBLEY, MD .

NEURAL NETWORKS, 1995, 8 (01) :11-23

← 1 2 →