Standard Article

Independent Component Analysis

  1. Riccardo Boscolo

Published Online: 14 APR 2006

DOI: 10.1002/9780471740360.ebs0622

Wiley Encyclopedia of Biomedical Engineering

Wiley Encyclopedia of Biomedical Engineering

How to Cite

Boscolo, R. 2006. Independent Component Analysis. Wiley Encyclopedia of Biomedical Engineering. .

Author Information

  1. University of California, Department of Electrical Engineering, Los Angeles, California

Publication History

  1. Published Online: 14 APR 2006

Abstract

The term Independent Component Analysis (ICA) broadly refers to the set of statistical principles as well as to the estimation algorithms derived from them, whose aim is to extract statistically independent components in the data, generally through some filtering procedure that can be linear or non-linear, act on a batch of data at the time or process the data samples as they become available (adaptive).

The field of Independent Component Analysis can be seen as stemming from the groundwork of several researchers, who in the last decade of the 20th century investigated fundamental statistical problems, such as the relationship between gaussianity and statistical dependence, redundancy reduction principles, and information preserving networks.

Keywords:

  • independent component analysis;
  • ICA;
  • Blind signal separation;
  • gaussianity;
  • information theory;
  • entropy;
  • DNA microarry analysis;
  • EEG