Sign and rank covariance matrices. (English) Zbl 0965.62049

Summary: The robust estimation of multivariate location and shape is one of the most challenging problems in statistics and crucial in many application areas. The objective is to find highly efficient, robust, computable and affine equivariant location and covariance matrix estimates. In this paper, three different concepts of multivariate sign and rank are considered and their ability to carry information about the geometry of the underlying distribution (or data cloud) is discussed. New techniques for robust covariance matrix estimation based on different sign and rank concepts are proposed and algorithms for computing them outlined.
In addition, new tools for evaluating the qualitative and quantitative robustness of a covariance estimator are proposed. The use of these tools is demonstrated on two rank-based covariance matrix estimates. Finally, to illustrate the practical importance of the problem, a signal processing example where robust covariance matrix estimates are needed is given.


62H12 Estimation in multivariate analysis
62G35 Nonparametric robustness
Full Text: DOI


[1] Bensmail, H.; Celeux, G., Regularized Gaussian discriminant analysis through eigenvalue decomposition, J. American statist. assoc., 91, 1743-1749, (1996) · Zbl 0885.62068
[2] Brown, B.M., Hettmansperger, T.P., Möttönen, J., Oja, H., 1997. Rank plots in the affine invariant case. In: Y. Dodge (Ed.), L1-Statistical Analysis and Related Methods. 351-362. · Zbl 0933.62004
[3] Campbell, N.A., Robust procedures in multivariate analysis I: robust covariance estimation., Appl. statist., 29, 231-237, (1980) · Zbl 0471.62047
[4] Capéraà, P.; Genest, C., Spearman’s ρ is larger than Kendall’s τ for positively dependent random variables, Nonparametric statist., 2, 183-194, (1993) · Zbl 1360.62294
[5] Cardoso, J.-F., 1989. Source separation using higher order moments. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP-89), 2109-2112.
[6] Chaudhuri, P., On a geometrical notion of quantiles for multivariate data, Journal of the American statistical association, 91, 862-872, (1996) · Zbl 0869.62040
[7] Choi, K.; Marden, J., A multivariate version of Kendall’s τ, Nonparametric statistics, 9, 261-293, (1998) · Zbl 0945.62062
[8] Croux, C., Ruiz-Gazen, A., 1996. A fast algorithm for robust principal components based on projection pursuit. In: A. Prat (Ed.), Compstat: Proc. Computational Statistics, Physica-Verlag, Heidelberg, pp. 211-217. · Zbl 0900.62300
[9] Croux, C., Haesbroeck, G., Rousseeuw, P., 1997. Location adjustment for the minimum volume ellipsoid estimator. Preprint University of Brussels (ULB). · Zbl 0900.62278
[10] Davies, P.L., Asymptotic behavior of S-estimators of multivariate location parameters and dispersion matrices, The annals of statistics, 15, 1269-1292, (1987) · Zbl 0645.62057
[11] Donoho, D.L., 1982. Breakdown properties of multivariate location estimators. Ph.D. Thesis, Harvard University.
[12] Gibbons, J.D.; Chakraborti, S., Nonparametric statistical inference., (1992), Marcel Dekker New York · Zbl 0785.62030
[13] Hampel, F.R.; Ronchetti, E.M.; Rousseeuw, P.J.; Stahel, W.A., Robust statistics., (1986), Wiley New York
[14] Hettmansperger, T.P.; Möttönen, J.; Oja, H., Multivariate affine invariant one-sample signed-rank tests, Journal of the American statistical society, 92, 1591-1600, (1997) · Zbl 0943.62051
[15] Hettmansperger, T.P.; Möttönen, J.; Oja, H., Affine invariant multivariate rank tests for several samples, Statistica sinica, 8, 785-800, (1998) · Zbl 0905.62062
[16] Hettmansperger, T.P.; Nyblom, J.; Oja, H., Affine invariant multivariate one-sample sign tests, Journal of the royal statistical society, series B, 56, 221-234, (1994) · Zbl 0795.62055
[17] Huber, P.J., 1977. Gupta, S.S., Moore, D.S. (Eds.), Statistical Decision Theory and Related Topics II. Academic Press, New York.
[18] Huber, P.J., Robust statistics., (1981), Wiley New York
[19] Kendall, M.G., Rank correlation methods., (1975), Griffin London · Zbl 0032.17602
[20] Kent, J.T.; Tyler, D.E., Redescending M-estimates of multivariate location and scatter, The annals of statistics, 19, 2102-2119, (1991) · Zbl 0763.62030
[21] Koivunen, V., Kassam, S.A., 1996. Covariance estimation in multivariate OS-filtering. IEEE ICIP-96, Vol. II, pp. 981-984.
[22] Koivunen, V., Pajunen, P., Karhunen, J., Oja, E., 1998. Blind separation from ε-contaminated mixtures. EUSIPCO’98, Vol. III, pp. 1817-1820.
[23] Koltchinskii, V., M-estimation, convexity and quantiles, The annals of statistics, 25, 435-477, (1997) · Zbl 0878.62037
[24] Li, G.; Chen, Z., Projection-pursuit approach to robust dispersion matrices and principal components: primary theory and Monte-Carlo, Journal of the American statistical association, 80, 759-766, (1985) · Zbl 0595.62060
[25] Lopuhaä, H.P., On the relation between S-estimators and M-estimators of multivariate location and covariance, The annals of statistics, 17, 1662-1683, (1989) · Zbl 0702.62031
[26] Lopuhaä, H.P.; Rousseeuw, P.J., Breakdown points of affine-equivariant estimators of multivariate location and covariance matrices, The annals of statistics, 19, 229-248, (1991) · Zbl 0733.62058
[27] Marden, J.I., 1999a. Multivariate rank tests. In: Ghosh (Ed.), Multivariate, Design, and Sampling, Marcel Dekker, New York. · Zbl 0946.62060
[28] Marden, J.I., 1999b. Some robust estimates of principal components. Statistics and Probability Letters 43, 349-359. · Zbl 0939.62055
[29] Maronna, R.A., Robust M-estimators of multivariate location and scatter, The annals of statistics, 20, 51-67, (1976) · Zbl 0322.62054
[30] Masarotto, G., Robust identification of autoregressive moving average models, Applied statistics, 36, 214-220, (1987) · Zbl 0617.62098
[31] Möttönen, J.; Oja, H., Multivariate spatial sign and rank methods, Journal of nonparametric statistics, 5, 201-213, (1995) · Zbl 0857.62056
[32] Oja, E., The nonlinear PCA learning rule in independent component analysis, Neurocomputing, 17, 25-45, (1997)
[33] Oja, H., Descriptive statistics for multivariate distributions, Statistics and probability letters, 1, 327-332, (1983) · Zbl 0517.62051
[34] Oja, H., 1999. Affine invariant multivariate sign and rank tests and corresponding estimates: a review. Scandinavian Journal of Statistics 26, 319-343. · Zbl 0938.62063
[35] Oja, H., Koivunen, V., 1998. Robust covariance matrix estimation with signal processing applications. In: The Yearbook of the Finnish Statistical Society 1997. Helsinki, 87-109.
[36] Rissanen, J., Modeling by shortest data description, Automatica, 14, 465-471, (1978) · Zbl 0418.93079
[37] Rousseeuw, P.J., Least Median of squares regression, Journal of American statistical association, 79, 871-880, (1984) · Zbl 0547.62046
[38] Rousseeuw, P.J.; Leroy, A., Robust regression and outlier detection., (1987), Wiley New York · Zbl 0711.62030
[39] Stahel, W.A., 1981. Breakdown of covariance estimators. Research Report 31, Fachgruppe für Statistik, E.T.H. Zürich.
[40] Stewart, G.W.; Sun, J., Matrix perturbation theory., (1990), Academic Press New York
[41] Wax, M.; Kailath, T., Detection of signals by information theoretic criteria, IEEE transactions on acoustics, speech, and signal processing (T-ASSP), 33, 387-392, (1985)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.