Research of independent component analysis

Author

Yu, Xiaoyuan ; Cheng, Xiaoyin ; Fu, Y. ; Zhou, J. ; Hao, H. ; Yang, Xu ; Huang, Heng ; Zhang, Tianzhu ; Fang, L.

Author_Institution

Inst. of Inf. Sci., Beijing Normal Univ.

Volume

5

fYear

0

fDate

0-0 0

Firstpage

4804

Abstract

Independent component analysis (ICA) is a statistical technique to decompose multivariate data into statistically independent components. It could be applied to mine data of medical, economy or telecommunication systems, and to analyze data of GIS systems for agriculture or environment applications. To solve the problem of blind source separation, this paper introduces the theory and developments of ICA. The analyses of different methods as well as the comparisons with each other on objective functions and optimization algorithms are given. Finally, some problems of ICA to be solved are discussed

Keywords

blind source separation; independent component analysis; optimisation; blind source separation; independent component analysis; multivariate data decomposition; optimization algorithms; Algorithm design and analysis; Blind source separation; Data mining; Feature extraction; Independent component analysis; Maximum likelihood estimation; Mutual information; Optimization methods; Principal component analysis; Signal processing algorithms;

fLanguage

English

Publisher

ieee

Conference_Titel

Systems, Man and Cybernetics, 2004 IEEE International Conference on

Conference_Location

The Hague

ISSN

1062-922X

Print_ISBN

0-7803-8566-7

Type

conf

DOI

10.1109/ICSMC.2004.1401291

Filename

1401291