## The $$k$$-step spatial sign covariance matrix.(English)Zbl 1284.62367

Summary: The sign covariance matrix is an orthogonal equivariant estimator of multivariate scale. It is often used as an easy-to-compute and highly robust estimator. In this paper we propose a $$k$$-step version of the sign covariance matrix, which improves its efficiency while keeping the maximal breakdown point. If $$k$$ tends to infinity, Tyler’s M-estimator is obtained. It turns out that even for very low values of $$k$$, one gets almost the same efficiency as Tyler’s M-estimator.

### MSC:

 62H30 Classification and discrimination; cluster analysis (statistical aspects) 62H25 Factor analysis and principal components; correspondence analysis 62H12 Estimation in multivariate analysis

MNM
Full Text:

### References:

 [1] Bensmail H, Celeux G (1996) Regularized Gaussian discriminant analysis through eigenvalue decomposition. J Am Stat Assoc 91: 1743–1749 · Zbl 0885.62068 [2] Croux C, Haesbroeck G (2000) Principal component analysis based on robust estimators of the covariance or correlation matrix: influence functions and efficiencies. Biometrika 87: 603–618 · Zbl 0956.62047 [3] Croux C, Ollila E, Oja H (2002) Sign and rank covariance matrices: statistical properties and application to principal components analysis. In: Dodge Y (eds) Statistical data analysis based on the L1-norm and related methods. Birkhauser, Basel, pp 257–271 · Zbl 1145.62343 [4] Dümbgen L, Tyler DE (2005) On the breakdown properties of some multivariate M-functionals. Scand J Stat 32: 247–264 · Zbl 1089.62056 [5] Frahm G (2009) Asymptotic distiributions of robust shape matrices and scale. J Multivar Anal 100: 1329–1337 · Zbl 1274.62130 [6] Hallin M, Oja H, Paindaveine D (2006) Semiparametrically efficient rank-based inference for shape. II. Optimal R-estimation of shape. Ann Stat 34: 2757–2789 · Zbl 1115.62059 [7] Hettmansperger TP, Randles RH (2002) A practical affine equivariant multivariate médian. Biometrika 89(4): 851–860 · Zbl 1036.62045 [8] Kent JT, Tyler DE (1991) Redescending M-estimates of multivariate location and scatter. Ann Stat 19: 2102–2119 · Zbl 0763.62030 [9] Locantore N, Marron JS, Simpson DG, Tripoli N, Zhang JT, Cohen KL (1999) Robust principal components for functional data. Test 8: 1–28 · Zbl 0980.62049 [10] Lopuhäa HP, Rousseeuw PJ (1991) Breakdown points of affine equivariant estimators of multivariate location and covariance matrices. Ann Stat 19: 229–248 · Zbl 0733.62058 [11] Maronna R, Martin D, Yohai V (2006) Robust statistics. Wiley, New York [12] Oja H (2010) Multivariate nonparametric methods with R. An approach based on spatial signs and ranks, Springer, Berlin (in press) · Zbl 1269.62036 [13] Ollila E, Heetmansperger TP, Oja H (2002) Affine equivariant multivariate sign methods (unpublished) [14] Paindaveine D (2008) A canonical definition of shape. Stat Probab Lett 78: 2240–2247 · Zbl 1283.62124 [15] Rousseeuw PJ, Croux C (1994) The bias of k-step M-estimators. Stat Probab Lett 20: 411–420 · Zbl 0801.62036 [16] Sirkia S, Taskinen S, Oja H, Tyler D (2009) Tests and estimates of shape based on spatial signs and ranks. J Nonparametr Stat 21: 15–176 · Zbl 1359.62150 [17] Taskinen S, Sirkiä S, Oja H (2010) k-step Shape estimators based on spatial signs and ranks. J Stat Plan Inference (in press) · Zbl 1372.62007 [18] Tyler DE (1987) A distribution-free M-estimator of multivariate scatter. Ann Stat 15: 234–251 · Zbl 0628.62053 [19] Visuri S, Ollila E, Koivunen V, Möttönen J, Oja H (2003) Affine equivariant multivariate rank methods. J Stat Plan Infer 114: 161–185 · Zbl 1011.62053 [20] Yadine A (2006) Robustness and efficiency of multivariate scatter estimators”, Ph.D. dissertation, Université Libre de Bruxelles, Bruxelles
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.