Expectation-maximization approaches to independent component analysis

被引：4

作者：

Zhong, MJ

Tang, HW

Tang, YY ^{[1
]}

机构：

[1] Dalian Univ Technol, Inst Neuroinformat, Dalian 116023, Peoples R China

[2] Dalian Univ Technol, Inst Computat Biol & Bioinformat, Dalian 116023, Peoples R China

[3] Chinese Acad Sci, Lab Visual Informat Proc, Beijing 100101, Peoples R China

[4] Chinese Acad Sci, Key Lab Mental Hlth, Beijing 100101, Peoples R China

来源：

NEUROCOMPUTING | 2004年 / 61卷

基金：

中国国家自然科学基金;

关键词：

independent component analysis; overcomplete representations; EM algorithm; variational method;

D O I：

10.1016/j.neucom.2004.06.007

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Expectation-Maximization (EM) algorithms for independent component analysis are presented in this paper. For super-Gaussian sources, a variational method is employed to develop an EM algorithm in closed form for learning the mixing matrix and inferring the independent components. For sub-Gaussian sources, a symmetrical form of the Pearson mixture model (Neural Comput. 11 (2) (1999) 417-441) is used as the prior, which also enables the development of an EM algorithm in fclosed form for parameter estimation. (C) 2004 Elsevier B.V. All rights reserved.

引用

页码：503 / 512

页数：10

共 13 条

[1] AN INFORMATION MAXIMIZATION APPROACH TO BLIND SEPARATION AND BLIND DECONVOLUTION [J].