Robust Independent Component Analysis Based on Two Scatter Matrices

Authors

  • Klaus Nordhausen University of Tampere, Finland
  • Hannu Oja University of Tampere, Finland
  • Esa Ollila Helsinki University of Technology, Finland

DOI:

https://doi.org/10.17713/ajs.v37i1.290

Abstract

Oja, Sirkiä, and Eriksson (2006) and Ollila, Oja, and Koivunen (2007) showed that, under general assumptions, any two scatter matrices with the so called independent components property can be used to estimate the unmixing matrix for the independent component analysis (ICA). The method is a generalization of Cardoso’s (Cardoso, 1989) FOBI estimate which uses the regular covariance matrix and a scatter matrix based on fourth moments. Different choices of the two scatter matrices are compared in a simulation study. Based on the study, we recommend always the use of two robust scatter matrices. For possible asymmetric independent components, symmetrized versions of the scatter matrix estimates should be used.

References

Amari, S., Cichocki, A., and Yang, H. (1996). A new learning algorithm for blind source separation. In Advances in neural information processing systems 8 (p. 757-763).

Cambridge, MA.: MIT Press.

Cardoso, J. (1989). Source separation using higher order moments. In Proceedings of IEEE international conference on acustics, speech and signal processing (p. 2109-

. Glasgow.

Dümbgen, L. (1998). On Tyler’s M-functional of scatter in high dimension. Annals of Institute of Statistical Mathematics, 50, 471-491.

Hettmansperger, T. P., and Randles, R. H. (2002). A practical affine equivariant multivariate median. Biometrika, 89, 851–860.

Hyvärinen, A., Karhunen, J., and Oja, E. (2001). Independent component analysis. New

York: Wiley.

Hyvärinen, A., and Oja, E. (1997). A fast fixed-point algorithm for independent component analysis. Neural Computation, 9, 1483-1492.

Hyvärinen, A., and Oja, E. (2000). Independent component analysis: Algorithms and applications. Neural Networks, 13, 411-430.

Marchini, J., Heaton, C., and Ripley, B. (2006). fastICA: FastICA algorithms to perform ICA and projection pursuit [Computer software manual]. (R package version 1.1-8)

Maronna, R., Martin, R., and Yohai, V. (2006). Robust statistics. Chichester: Wiley.

Nordhausen, K., Oja, H., and Tyler, D. (2006). ICS: ICS / ICA computation based on two scatter matrices [Computer software manual]. (R package version 0.1-2)

Oja, H., Sirkiä, S., and Eriksson, J. (2006). Scatter matrices and independent component analysis. Austrian Journal of Statistics, 35, 175-189.

Ollila, E., Oja, H., and Koivunen, V. (2007). Complex-valued ICA based on a pair of generalized covariance matrices. (Conditionally accepted by Computational Statistics & Data Analysis)

R Development Core Team. (2006). R: A language and environment for statistical computing [Computer software manual]. Vienna, Austria. Available from

http://www.R-project.org (ISBN 3-900051-07-0)

Sirkiä, S., Taskinen, S., and Oja, H. (2007). Symmetrized M-estimators of multivariate scatter. Journal of Multivariate Analysis, 98, 1611-1629.

Tyler, D. E. (1987). A distribution-free M-estimator of multivariate scatter. Annals of Statistics, 15, 234-251.

Downloads

Published

2016-04-03

How to Cite

Nordhausen, K., Oja, H., & Ollila, E. (2016). Robust Independent Component Analysis Based on Two Scatter Matrices. Austrian Journal of Statistics, 37(1), 91–100. https://doi.org/10.17713/ajs.v37i1.290

Issue

Section

Articles