Abstract
We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminish reduce its effects.
Keywords
- Independent Component Analysis
- Blind Source Separation
- Flicker Noise
- Singular Spectrum Analysis
- Source Estimate
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
P. Comon. Independent component analysis-a new concept? Signal Processing, 36:287–314, 1994.
The FastICA MATLAB package. Available at http://www.cis.hut.fi/projects/ica/fastica/, 1998.
M. Hämäläinen, R. Hari, R.J. Ilmoniemi, J. Knuutila, and O.V. Lounasmaa. Magnetoencephalography-theory, instrumentation, and applications to noninvasive studies of the working human brain. Reviews of Modern Physics, 65:413–497, 1993.
A. Hyvärinen, J. Karhunen, and E. Oja. Independent component analysis. Wiley, 2001. in press.
A. Hyvärinen, J. Särelä, and R. Vigário. Spikes and bumps: Artefacts generated by independent component analysis with insufficient sample size. In Proc. Int. Workshop on Independent Component Analysis and Blind Signal Separation (ICA'99), Aussois, France, 1999. To appear.
0. Jahn and A. Cichocki. Identification and elimination of artifacts from MEG signals using efficient independent components analysis. In Proc. of the 11th Int. Conf. on Biomagnetism (BIOMAG-98), Sendai, Japan, 1998.
C. Jutten and J. Herault. Blind separation of sources, part I: An adaptive algorithm based on neuromimetic architecture. Signal Processing, 24:1–10, 1991.
Proves of the propositions in the paper. Available at http://www.cis.hut.fi/jaakko/proves/IWANNO1.ps, 2001.
R. Vigário, J. Särelä, V. Jousmäki, M. Hämäläinen, and E. Oja. Independent component approach to the analysis of eeg and meg recordings. IEEE transactions on biomedical engineering, 47(5):589–593, 2000.
R.N. Vigário, V. Jousmäki, M. Hämäläinen, R. Hari, and E. Oja. Independent component analysis for identification of artifacts in magnetoencephalographics recordings. In Advances in Neural Information Processing 10 (Proc. NIPSi97), pages 229–235. MIT Press, Cambridge MA, 1997.
Pascal Yiou, Didier Sornette, and Michael Ghil. Data-adaptive wavelets and multiscale singular-spectrum analysis. Physica D, 142:254–290, 2000.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2001 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Särelä, J., Vigário, R. (2001). The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions. In: Mira, J., Prieto, A. (eds) Bio-Inspired Applications of Connectionism. IWANN 2001. Lecture Notes in Computer Science, vol 2085. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45723-2_99
Download citation
DOI: https://doi.org/10.1007/3-540-45723-2_99
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-42237-2
Online ISBN: 978-3-540-45723-7
eBook Packages: Springer Book Archive