Abstract
Most ICA algorithms are sensitive to outliers. Instead of robustifying existing algorithms by outlier rejection techniques, we show how a simple outlier index can be used directly to solve the ICA problem for super-Gaussian source signals. This ICA method is outlier-robust by construction and can be used for standard ICA as well as for overcomplete ICA (i.e. more source signals than observed signals (mixtures)).
| Original language | English |
|---|---|
| Pages (from-to) | 217-224 |
| Number of pages | 8 |
| Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Volume | 3195 |
| Publication status | Published - 2004 Dec 1 |
| Externally published | Yes |
Fingerprint
ASJC Scopus subject areas
- Computer Science(all)
- Biochemistry, Genetics and Molecular Biology(all)
- Theoretical Computer Science
Cite this
Robust ICA for super-Gaussian sources. / Meinecke, Frank C.; Harmeling, Stefan; Muller, Klaus.
In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 3195, 01.12.2004, p. 217-224.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Robust ICA for super-Gaussian sources
AU - Meinecke, Frank C.
AU - Harmeling, Stefan
AU - Muller, Klaus
PY - 2004/12/1
Y1 - 2004/12/1
N2 - Most ICA algorithms are sensitive to outliers. Instead of robustifying existing algorithms by outlier rejection techniques, we show how a simple outlier index can be used directly to solve the ICA problem for super-Gaussian source signals. This ICA method is outlier-robust by construction and can be used for standard ICA as well as for overcomplete ICA (i.e. more source signals than observed signals (mixtures)).
AB - Most ICA algorithms are sensitive to outliers. Instead of robustifying existing algorithms by outlier rejection techniques, we show how a simple outlier index can be used directly to solve the ICA problem for super-Gaussian source signals. This ICA method is outlier-robust by construction and can be used for standard ICA as well as for overcomplete ICA (i.e. more source signals than observed signals (mixtures)).
UR - http://www.scopus.com/inward/record.url?scp=35048843771&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=35048843771&partnerID=8YFLogxK
M3 - Article
VL - 3195
SP - 217
EP - 224
JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SN - 0302-9743
ER -