Abstract
Dimension reduction is a common strategy in multivariate data analysis which seeks a subspace which contains all interesting features needed for the subsequent analysis. Non-Gaussian component analysis attempts for this purpose to divide the data into a non-Gaussian part, the signal, and a Gaussian part, the noise. We will show that the simultaneous use of two scatter functionals can be used for this purpose and suggest a bootstrap test to test the dimension of the non-Gaussian subspace. Sequential application of the test can then for example be used to estimate the signal dimension.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Archimbaud, A., Nordhausen, K., Ruiz-Gazen, A.: ICS for multivariate outlier detection with application to quality control. Comput. Statistics Data Anal. 128, 184–199 (2018)
Bean, D.M.: Non-Gaussian component analysis. Ph.D. thesis, University of California, Berkeley (2014)
Blanchard, G., Kawanabe, M., Sugiyama, M., Spokoiny, V., Müller, K.-R.: In search of non-Gaussian components of a high-dimensional distribution. J. Mach. Learn. Res. 7, 247–282 (2006)
Blanchard, G., Sugiyama, M., Kawanabe, M., Spokoiny, V., Müller, K.-R.: Non-Gaussian component analysis: a semi-parametric framework for linear dimension reduction. In: Advances in Neural Information Processing Systems, pp. 131–138 (2005)
Cardoso, J.-F.: Source separation using higher order moments. In: International Conference on Acoustics, Speech, and Signal Processing, pp. 2109–2112 (1989)
Comon, P., Jutten, C.: Handbook of Blind Source Separation: Independent Component Analysis and Applications. Academic Press, Amsterdam (2010)
Dümbgen, L., Nordhausen, K., Schuhmacher, H.: New algorithms for M-estimation of multivariate scatter and location. J. Multivariate Anal. 144, 200–217 (2016)
Dümbgen, L., Pauly, M., Schweizer, T.: M-functionals of multivariate scatter. Statistics Surv. 9, 32–105 (2015)
Fischer, D., Berro, A., Nordhausen, K., Ruiz-Gazen, A.: REPPlab: an R package for detecting clusters and outliers using exploratory projection pursuit. In: Communications in Statistics—Simulation and Computation, pp. 1–23 (2019)
Huber, P.J.: Robust estimation of a location parameter. Ann Math Statistics 35:73–101 (1964)
Huber, P.J.: Projection pursuit. Ann. Statistics 13, 435–475 (1985)
Jin, Z., Risk, B.B., Matteson, D.S.: Optimization and testing in linear non-Gaussian component analysis. Statistical Anal. Data Mining ASA Data Sci. J. 12, 141–156 (2019)
Jolliffe, I.T.: Principal Component Analysis, 2nd edn. Springer, New York (2002)
Jones, M.C., Sibson, R.: What is projection pursuit? J. R. Statistical Soc. Ser. A 150, 1–37 (1987)
Kawanabe, M., Sugiyama, M., Blanchard, G., Müller, K.-R.: A new algorithm of non-Gaussian component analysis with radial kernel functions. Ann. Inst. Statistical Math. 59, 57–75 (2007)
Kent, J.T., Tyler, D.E.: Redescending M-estimates of multivariate location and scatter. Ann. Statistics 19, 2102–2119 (1991)
Miettinen, J., Nordhausen, K., Taskinen, S.: Blind source separation based on joint diagonalization in R: the packages JADE and BSSasymp. J. Statistical Softw. 76, 1–31 (2017)
Miettinen, J., Nordhausen, K., Taskinen, S., Tyler, D.E.: On the computation of symmetrized M-estimators of scatter. In: Agostinelli, C., Basu, A., Filzmoser, P., Mukherjee, D. (eds.) Recent Advances in Robust Statistics: Theory and Applications, pp. 151–167. Springer, New Delhi (2016)
Miettinen, J., Taskinen, S., Nordhausen, K., Oja, H.: Fourth moments and independent component analysis. Statistical Sci. 30, 372–390 (2015)
Nordhausen, K., Oja, H.: Independent subspace analysis using three scatter matrices. Austr. J. Statistics 40, 93–101 (2016)
Nordhausen, K., Oja, H.: Independent component analysis: a statistical perspective. Wiley Interdiscip. Rev. Comput. Statistics 10, e1440 (2018)
Nordhausen, K., Oja, H., Ollila, E.: Multivariate models and the first four moments. In: Hunter, D.R., Richards, D.S.R., Rosenberger, J.L. (eds.) Nonparametric Statistics and Mixture Models: A Festschrift in Honour of Thomas P. Hettmansperger, pp. 267–287. World Scientific, Singapore (2011)
Nordhausen, K., Oja, H., Ollila, E.: Robust independent component analysis based on two scatter matrices. Austr. J. Statistics 37, 91–100 (2016)
Nordhausen, K., Oja, H., Tyler, D.E.: Tools for exploring multivariate data: The package ICS. J. Statistical Softw. 28, 1–31 (2008)
Nordhausen, K., Oja, H., Tyler, D.E., Virta, J.: Asymptotic and bootstrap tests for the dimension of the non-Gaussian subspace. IEEE Signal Process. Lett. 24, 887–891 (2017)
Nordhausen, K., Oja, H., Tyler, D.E., Virta, J.: ICtest: estimating and testing the number of interesting components in linear dimension reduction. R package version 0.3-2 (2019)
Nordhausen, K., Oja, H., Tyler, D.E.: Asymptotic and bootstrap tests for subspace dimension. arXiv:1611.04908 (2017)
Nordhausen, K., Tyler, D.E.: A cautionary note on robust covariance plug-in methods. Biometrika 102, 573–588 (2015)
Nordhausen, K., Virta, J.: An overview of properties and extensions of FOBI. Knowl.-Based Syst. 173, 113–116 (2019)
Oja, H., Sirkiä, S., Eriksson, J.: Scatter matrices and independent component analysis. Austr. J. Statistics 35, 175–189 (2006)
R Core Team.: R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria (2017)
Risk, Benjamin B., Matteson, David S., Ruppert, David: Linear non-Gaussian component analysis via maximum likelihood. J. Am. Statistical Assoc. 114, 332–343 (2019)
Sasaki, H., Niu, G., Sugiyama, M.: Non-Gaussian component analysis with log-density gradient estimation. In: Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, pp. 1177–1185 (2016)
Sirkiä, S., Miettinen, J., Nordhausen, K., Oja, H., Taskinen, S.: SpatialNP: multivariate nonparametric methods based on spatial signs and ranks. R package version 1.1-4 (2019)
Sirkiä, S., Taskinen, S., Oja, H.: Symmetrised M-estimators of multivariate scatter. J. Multivariate Anal. 98, 1611–1629 (2007)
Theis, F.J.: Towards a general independent subspace analysis. In: Schölkopf, B., Platt, J.C., Hoffman, T. (eds.) Advances in Neural Information Processing Systems 19, pp. 1361–1368. MIT Press, Cambridge, MA (2007)
Theis, F.J., Kawanabe, M., Müller, K.R.: Uniqueness of non-Gaussianity-based dimension reduction. IEEE Trans. Signal Process. 59(9), 4478–4482 (2011)
Tyler, D., Critchley, F., Dümbgen, L., Oja, H.: Invariant coordinate selection. J. R. Statistical Soc. Ser. B 71, 549–592 (2009)
Urbanek, S.: png: read and write PNG images. R package version 0.1-7 (2013)
Ushey, K.: RcppRoll: efficient rolling/windowed operations. R package version 0.3-0 (2018)
Virta, J., Nordhausen, K., Oja, H.: Projection pursuit for non-Gaussian independent components. arXiv preprint arXiv:1612.05445 (2016)
Wolodzko, T.: extraDistr: additional univariate and multivariate distributions. R package version 1.8-11 (2019)
Acknowledgements
The authors wish to express their gratitude to two anonymous referees, whose insightful comments greatly helped improving the quality of the manuscript. The work of KN was supported by the Austrian Science Fund (FWF) Grant number P31881-N32.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Radojičić, U., Nordhausen, K. (2020). Non-Gaussian Component Analysis: Testing the Dimension of the Signal Subspace. In: Maciak, M., Pešta, M., Schindler, M. (eds) Analytical Methods in Statistics. AMISTAT 2019. Springer Proceedings in Mathematics & Statistics, vol 329. Springer, Cham. https://doi.org/10.1007/978-3-030-48814-7_6
Download citation
DOI: https://doi.org/10.1007/978-3-030-48814-7_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-48813-0
Online ISBN: 978-3-030-48814-7
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)