TY - JOUR
T1 - A resampling approach to estimate the stability of one-dimensional or multidimensional independent components
AU - Meinecke, Frank
AU - Ziehe, Andreas
AU - Kawanabe, Motoaki
AU - Müller, Klaus Robert
N1 - Funding Information:
Manuscript received December 20, 2001; revised June 3, 2002. The work of K.-R. Müller and A. Ziehe was supported in part by the EU Project (IST-1999-14190—BLISS). The studies were supported by the Bundesmin-isterium für Bildung und Forschung (BMBF) under Grant FKZ 01IBB02A and Grant 01IBB02B. This work is an extension of previous conference publications. Asterisk indicates corresponding author.
PY - 2002/12/1
Y1 - 2002/12/1
N2 - When applying unsupervised learning techniques in biomedical data analysis, a key question is whether the estimated parameters of the studied system are reliable. In other words, can we assess the quality of the result produced by our learning technique? We propose resampling methods to tackle this question and illustrate their usefulness for blind-source separation (BSS). We demonstrate that our proposed reliability estimation can be used to discover stable one-dimensional or multidimensional independent components, to choose the appropriate BSS-model, to enhance significantly the separation performance, and, most importantly, to flag components that carry physical meaning. Application to different biomedical testbed data sets (magnetoencephalography (MEG)/electrocardiography (ECG)-recordings) underline the usefulness of our approach.
AB - When applying unsupervised learning techniques in biomedical data analysis, a key question is whether the estimated parameters of the studied system are reliable. In other words, can we assess the quality of the result produced by our learning technique? We propose resampling methods to tackle this question and illustrate their usefulness for blind-source separation (BSS). We demonstrate that our proposed reliability estimation can be used to discover stable one-dimensional or multidimensional independent components, to choose the appropriate BSS-model, to enhance significantly the separation performance, and, most importantly, to flag components that carry physical meaning. Application to different biomedical testbed data sets (magnetoencephalography (MEG)/electrocardiography (ECG)-recordings) underline the usefulness of our approach.
KW - Blind-source separation
KW - Bootstrap
KW - Electrocardiography (ECG)
KW - Independent component analysis
KW - Magnetoencephalography (MEG)
KW - Multidimensional independent component analysis (ICA)
KW - Reliability
KW - Resampling
KW - Stability
KW - Unsupervised learning
UR - http://www.scopus.com/inward/record.url?scp=0036928393&partnerID=8YFLogxK
U2 - 10.1109/TBME.2002.805480
DO - 10.1109/TBME.2002.805480
M3 - Article
C2 - 12549733
AN - SCOPUS:0036928393
SN - 0018-9294
VL - 49
SP - 1514
EP - 1525
JO - IEEE Transactions on Biomedical Engineering
JF - IEEE Transactions on Biomedical Engineering
IS - 12 I
ER -