Sankhya B

, Volume 72, Issue 2, pp 123–153 | Cite as

Projection pursuit via white noise matrices



Projection pursuit is a technique for locating projections from high- to low-dimensional space that reveal interesting non-linear features of a data set, such as clustering and outliers. The two key components of projection pursuit are the chosen measure of interesting features (the projection index) and its algorithm. In this paper, a white noise matrix based on the Fisher information matrix is proposed for use as the projection index. This matrix index is easily estimated by the kernel method. The eigenanalysis of the estimated matrix index provides a set of solution projections that are most similar to white noise. Application to simulated data and real data sets shows that our algorithm successfully reveals interesting features in fairly high dimensions with a practical sample size and low computational effort.


Projection pursuit Fisher information matrix Eigenanalysis 


  1. Ahn, J., J. Marron, K. Muller, and Y. Chi 2007. The high-dimension, low-sample-size geometric representation holds under mild conditions. Biometrika 94(3):760.MATHCrossRefMathSciNetGoogle Scholar
  2. Azzalini, A., and A. Capitanio 1999. Statistical applications of the multivariate skew normal distribution. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 61(3):579–602.MATHCrossRefMathSciNetGoogle Scholar
  3. Azzalini, A., and A. Capitanio 2003. Distributions generated by perturbation of symmetry with emphasis on a multivariate skew t-distribution. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 65(2):367–389.MATHCrossRefMathSciNetGoogle Scholar
  4. Azzalini, A., and A. Valle 1996. The multivariate skew-normal distribution. Biometrika 83(4):715.MATHCrossRefMathSciNetGoogle Scholar
  5. Ballam, J., G.B. Chadwick, Z.C.G. Guiragossian, W.B. Johnson, D.W.G.S. Leith, and J. Morigasu. 1971. Van Hove analysis of the reactions π p → π π π + p and π + p → π + π + π at 16 GeV/C. Physical Review 4:1946–1947.Google Scholar
  6. Bowman, A.W., and P.J. Foster. 1993. Adaptive smoothing and density-based teste of multivariate normality. Journal of American Statistical Association 88(422):529–539.MATHCrossRefMathSciNetGoogle Scholar
  7. Calo, D.G. 2007. Gaussian mixture model classification: a projection pursuit approach. Computational Statistics & Data Analysis 52(1):471–482.MATHCrossRefMathSciNetGoogle Scholar
  8. Davison, A.C., and D.V. Hinkley. 1997. Bootstrap methods and their application. Cambridge Series in Statistical and Probabilistic Mathematics, No 1. ISBN-10: 0521574714.Google Scholar
  9. Diaconis, P., and D. Freedman. 1984. Asymptotics of graphical projection pursuit. Annals of Statistics 12(3):793–815.MATHCrossRefMathSciNetGoogle Scholar
  10. Fraley, C., and A. Raftery 2002. Model-based clustering, discriminant analysis, and density estimation. Journal of the American Statistical Association 97(458):611–631.MATHCrossRefMathSciNetGoogle Scholar
  11. Friedman, J.H. 1987. Exploratory projection pursuit. Journal of the American Statistical Association 82(397):249–266.MATHCrossRefMathSciNetGoogle Scholar
  12. Friedman, J.H., and J.W. Tukey. 1974. A projection pursuit algorithm for exploatory data analysis. IEEE Transactions on Computers C-23:881–889.CrossRefGoogle Scholar
  13. Frühwirth-Schnatter, S., and S. Pyne 2010. Bayesian inference for finite mixtures of univariate and multivariate skew-normal and skew-t distributions. Biostatistics 11(2):317.CrossRefGoogle Scholar
  14. Genton, M. 2004. Skew-elliptical distributions and their applications: a journey beyond normality.Google Scholar
  15. Godambe, V.P. 1960. An opertimal property of regular maximal likelihood estimation. Annals of Mathematical Statistics 31(4):1208–1211.CrossRefMathSciNetGoogle Scholar
  16. Golub, T., D. Slonim, P. Tamayo, C. Huard, M. Gaasenbeek, J. Mesirov, H. Coller, M. Loh, J. Downing, M. Caligiuri, et al. 1999. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 286(5439):531.CrossRefGoogle Scholar
  17. Hall, P., J.S. Marron, and A. Neeman. 2005. Geometric representation of high dimension, low sample size data. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 67(3):427–444.MATHCrossRefMathSciNetGoogle Scholar
  18. Huber, P.J. 1985. Projection pursuit. Annals of Statistics 13(2):435–475.MATHCrossRefMathSciNetGoogle Scholar
  19. Hui, G.D. 2008. Matrix distances with their application to finding directional deviations from normality in high-dimensinal data. PhD Thesis, Pennsylvania State University.Google Scholar
  20. Jee, R.J. 1985. A study of projection pursuit methods. PhD Thesis, Rice University.Google Scholar
  21. Kagan, A. 2001. Aother look at Cramer–Rao inequality. The American Statistician 55(3):211–212(2).MATHCrossRefMathSciNetGoogle Scholar
  22. Kagan, A., and Yu.V. Linnik, and C.R. Rao. 1973. Characterization problems in mathematical statistics. Wiley Series in Probability and Mathematical Statistics, No 1. ISBN-10: 0471454214Google Scholar
  23. Kazuyoshi, Y., and A. Makoto. 2001. Effective PCA for high-dimension, low-sample-size data with singular value decomposition of cross data matrix. Journal of Multivariate Analysis 101(9):2060–2077.Google Scholar
  24. Kazuyoshi, Y., and A. Makoto. 2009. PCA consistency for non-Gaussian data in high dimension, low sample size context. Communications in Statistics - Theory and Methods 38(16):2634–2652.MATHCrossRefMathSciNetGoogle Scholar
  25. Li, J., S. Ray, and B. Lindsay 2007. A nonparametric statistical approach to clustering via mode identification. Journal of Machine Learning Research 8(8):1687–1723.MathSciNetGoogle Scholar
  26. Lin, T., J. Lee, and S. Yen 2007. Finite mixture modelling using the skew normal distribution. Statistica Sinica 17(3):909.MATHMathSciNetGoogle Scholar
  27. Lindsay, B.G. 1982. Conditional score functions: some optimality results. Biometrika 69:503–512.MATHCrossRefMathSciNetGoogle Scholar
  28. Lindsay, B.G., M. Markatou, S.R. Ray, K. Yang, and S.C. Chen. 2008. Quadratic distances on probabilities: a unified foundation. Annals of Statistics 36:983–1006.MATHCrossRefMathSciNetGoogle Scholar
  29. Melnykov, V., R. and Maitra 2010. Finite mixture models and model-based clustering. Statistics Surveys 4:80–116.MATHCrossRefMathSciNetGoogle Scholar
  30. Muller, K.E., Y.-Y. Chi, J. Ahn, and J.S. Marron. 2011. Limitations of high dimension, low sample size principal components for Gaussian data (under revision for resubmission).Google Scholar
  31. Papaioannou, T., and K. Ferentinos. 2005. On two forms of Fisher’s measure of information. Communications in Statistics - Theory and Methods 34:1461–1470.MATHCrossRefMathSciNetGoogle Scholar
  32. Posse, C. 1995. Projection pursuit exploratory data analysis. Computational Statistics and Data Analysis 20:669–687.MATHCrossRefMathSciNetGoogle Scholar
  33. Ray, S., and B.G. Lindsay 2008. Model selection in high dimensions: a quadratic-risk-based approach. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 70(1):95–118.MATHMathSciNetGoogle Scholar
  34. Sungkyu, J., and J.S. Marron. 1995. PCA consistency in high dimension, low sample size context. Annals of Statistics 37(6B):4104–4130.Google Scholar
  35. Terrell, G.R. 1995. A Fisher information test for Pearson-family membership. In Proceedings of the statistical computing section, joint statistical meetings, Orlando, Florida, 230–234.Google Scholar

Copyright information

© Indian Statistical Institute 2011

Authors and Affiliations

  1. 1.Genzyme CorporationFraminghamUSA
  2. 2.Department of StatisticsPennsylvania State UniversityUniversity ParkUSA

Personalised recommendations