Copyright © 2006 The Institute of Electronics, Information and Communication Engineers
Regular Section -- Letters -- Digital Signal Processing |
An Extension to the Natural Gradient Algorithm for Robust Independent Component Analysis in the Presence of Outliers*
1 The authors are with the Department of Electronic Engineering, Tohoku University, Sendai-shi, 980-8579 Japan. E-mail: tufail{at}mk.ecei.tohoku.ac.jp
In this paper, we propose to employ an extension to the natural gradient algorithm for robust Independent Component Analysis against outliers. The standard natural gradient algorithm does not exhibit this property since it employs nonrobust sample estimates for computing higher order moments. In order to overcome this drawback, we propose to use robust alternatives to higher order moments, which are comparatively less sensitive to outliers in the observed data. Some computer simulations are presented to show that the proposed method, as compared to the standard natural gradient algorithm, gives better performance in the presence of outlying data.
Key Words: Independent Component Analysis, robust higher order statistics, outliers
Manuscript received January 25, 2006. Manuscript revised April 19, 2006. Final manuscript received June 1, 2006.
* This work was supported by the Monbukagakusho Government of Japan.