TY - GEN
T1 - ICA with sparse connections
T2 - 8th International Conference on Independent Component Analysis and Signal Separation, ICA 2009
AU - Zhang, Kun
AU - Peng, Heng
AU - Chan, Laiwan
AU - Hyvarinen, Aapo
N1 - Copyright:
Copyright 2012 Elsevier B.V., All rights reserved.
PY - 2009/2/25
Y1 - 2009/2/25
N2 - When applying independent component analysis (ICA), sometimes we expect the connections between the observed mixtures and the recovered independent components (or the original sources) to be sparse, to make the interpretation easier or to reduce the random effect in the results. In this paper we propose two methods to tackle this problem. One is based on adaptive Lasso, which exploits the L1 penalty with data-adaptive weights. We show the relationship between this method and the classic information criteria such as BIC and AIC. The other is based on optimal brain surgeon, and we show how its stopping criterion is related to the information criteria. This method produces the solution path of the transformation matrix, with different number of zero entries. These methods involve low computational loads. Moreover, in each method, the parameter controlling the sparsity level of the transformation matrix has clear interpretations. By setting such parameters to certain values, the results of the proposed methods are consistent with those produced by classic information criteria.
AB - When applying independent component analysis (ICA), sometimes we expect the connections between the observed mixtures and the recovered independent components (or the original sources) to be sparse, to make the interpretation easier or to reduce the random effect in the results. In this paper we propose two methods to tackle this problem. One is based on adaptive Lasso, which exploits the L1 penalty with data-adaptive weights. We show the relationship between this method and the classic information criteria such as BIC and AIC. The other is based on optimal brain surgeon, and we show how its stopping criterion is related to the information criteria. This method produces the solution path of the transformation matrix, with different number of zero entries. These methods involve low computational loads. Moreover, in each method, the parameter controlling the sparsity level of the transformation matrix has clear interpretations. By setting such parameters to certain values, the results of the proposed methods are consistent with those produced by classic information criteria.
KW - Independent Component Analysis
KW - Sparsity Level
KW - Adaptive Lasso
KW - Oracle Property
UR - http://www.scopus.com/inward/record.url?scp=67149131673&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-00599-2_25
DO - 10.1007/978-3-642-00599-2_25
M3 - Conference proceeding
AN - SCOPUS:67149131673
SN - 9783642005985
T3 - Lecture Notes in Computer Science
SP - 195
EP - 202
BT - Independent Component Analysis and Signal Separation
A2 - Adali, Tülay
A2 - Jutten, Christian
A2 - Romano, João Marcos Travassos
A2 - Barros, Allan Kardec
PB - Springer
CY - Berlin, Heidelberg
Y2 - 15 March 2009 through 18 March 2009
ER -