Abstract
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 L 1 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.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Akaike, H.: Information theory and an extension of the maximum likelihood principle. In: Proc. 2nd Int. Symp. on Information Theory, pp. 267–281 (1973)
Fan, J., Li, R.: Variable selection via nonconcave penalized likelihood and its oracle properties. J. Amer. Statist. Assoc. 96, 1348–1360 (2001)
Hassibi, B., Stork, D.G.: Second order derivatives for network pruning: Optimal brain surgeon. In: NIPS 5, pp. 164–171. Morgan Kaufmann, San Francisco (1993)
Hyvärinen, A., Karhunen, J., Oja, E.: Independent Component Analysis. John Wiley & Sons, Inc., Chichester (2001)
Hyvärinen, A., Karthikesh, R.: Imposing sparsity on the mixing matrix in independent component analysis. Neurocomputing 49, 151–162 (2002)
Pham, D.T., Garat, P.: Blind separation of mixture of independent sources through a quasi-maximum likelihood approach. IEEE Trans. on Signal Processing 45(7), 1712–1725 (1997)
Schwarz, G.: Estimating the dimension of a model. The Annals of Statistics 6, 461–464 (1978)
Shimizu, S., Hoyer, P.O., Hyvärinen, A., Kerminen, A.J.: A linear non-Gaussian acyclic model for causal discovery. JMLR 7, 2003–2030 (2006)
Silva, F.M., Almeida, L.B.: Acceleration techniques for the backpropagation algorithm. In: Neural Networks, pp. 110–119. Springer, Heidelberg (1990)
Tibshirani, R.: Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society 58(1), 267–288 (1996)
Zhang, K., Chan, L.-W.: ICA with sparse connections. In: Corchado, E., Yin, H., Botti, V., Fyfe, C. (eds.) IDEAL 2006. LNCS, vol. 4224, pp. 530–537. Springer, Heidelberg (2006)
Zhang, K., Chan, L.: Minimal nonlinear distortion principle for nonlinear independent component analysis. JMLR 9, 2455–2487 (2008)
Zhao, P., Yu, B.: On model selection consistency of lasso. JMLR 7, 2541–2563 (2006)
Zou, H.: The adaptive lasso and its oracle properties. Journal of the American Statistical Association 101(476), 1417–1429 (2006)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2009 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Zhang, K., Peng, H., Chan, L., Hyvärinen, A. (2009). ICA with Sparse Connections: Revisited. In: Adali, T., Jutten, C., Romano, J.M.T., Barros, A.K. (eds) Independent Component Analysis and Signal Separation. ICA 2009. Lecture Notes in Computer Science, vol 5441. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-00599-2_25
Download citation
DOI: https://doi.org/10.1007/978-3-642-00599-2_25
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-00598-5
Online ISBN: 978-3-642-00599-2
eBook Packages: Computer ScienceComputer Science (R0)