zbMATH — the first resource for mathematics

A scatter matrix estimate based on the zonotope. (English) Zbl 1046.62058
Summary: We introduce a new scatter matrix functional which is a multivariate affine equivariant extension of the mean deviation \(E(| x-\text{Med}(x) |)\). The estimate is constructed using the data vectors (centered with the multivariate Oja median) and their angular distances. The angular distance is based on the Randles interdirections [R. H. Randles, J. Am. Stat. Assoc. 84, No. 408, 1045–1050 (1989; Zbl 0702.62039)]. The new estimate is called the zonoid covariance matrix (the ZCM), as it is the regular covariance matrix of the centers of the facets of the zonotope based on the data set.
There is a kind of symmetry between the zonoid covariance matrix and the affine equivariant sign covariance matrix; interchanging the roles of data vectors and hyperplanes yields the sign covariance matrix as the zonoid covariance matrix. (It turns out that the symmetry relies on the zonoid of the distribution and its projection body which is also a zonoid.) The influence function and limiting distribution of the new scatter estimate, the ZCM, are derived to consider the robustness and efficiency properties of the estimate.
Finite-sample efficiencies are studied in a small simulation study. The influence function of the ZCM is unbounded (linear in the radius of the contamination vector) but less influential in the tails than that of the regular covariance matrix (quadratic in the radius). The estimate is highly efficient in the multivariate normal case and performs better than the regular covariance matrix for heavy-tailed distributions.

62H12 Estimation in multivariate analysis
62H05 Characterization and structure theory for multivariate probability distributions; copulas
Full Text: DOI
[1] Arcones, M. A., Chen, Z. and Giné, E. (1994). Estimators related to \(U\)-processes with applications to multivariate medians: Asymptotic normality. Ann. Statist. 22 1460–1477. JSTOR: · Zbl 0827.62023 · doi:10.1214/aos/1176325637 · links.jstor.org
[2] Croux, C. and Haesbroeck, G. (2000). Principal component analysis based on robust estimators of the covariance or correlation matrix: Influence functions and efficiencies. Biometrika 87 603–618. JSTOR: · Zbl 0956.62047 · doi:10.1093/biomet/87.3.603 · links.jstor.org
[3] Croux, C., Ollila, E. and Oja, H. (2002). Sign and rank covariance matrices: Statistical properties and application to principal component analysis. In Statistical Data Analysis Based on the \(L_1\) Norm and Related Methods (Y. Dodge, ed.) 257–270. Birkhäuser, Basel. · Zbl 1145.62343
[4] Gardner, R. J. (1995). Geometric Tomography. Cambridge Univ. Press. · Zbl 1042.52501
[5] Hallin, M. and Paindaveine, D. (2002). Optimal tests for multivariate location based on interdirections and pseudo-Mahalanobis ranks. Ann. Statist. 30 1103–1133. · Zbl 1101.62348 · doi:10.1214/aos/1031689019 · euclid:aos/1031689019
[6] Hampel, F. R., Ronchetti, E. M., Rousseeuw, P. J. and Stahel, W. A. (1986). Robust Statistics. The Approach Based on Influence Functions. Wiley, New York. · Zbl 0593.62027
[7] Koshevoy, G. and Mosler, K. (1997a). Multivariate Gini indices. J. Multivariate Anal. 60 252–276. · Zbl 0873.62062 · doi:10.1006/jmva.1996.1655
[8] Koshevoy, G. and Mosler, K. (1997b). Zonoid trimming for multivariate distributions. Ann. Statist. 25 1998–2017. · Zbl 0881.62059 · doi:10.1214/aos/1069362382
[9] Koshevoy, G. and Mosler, K. (1998). Lift zonoids, random convex hulls and the variability of random vectors. Bernoulli 4 377–399. · Zbl 0945.52006 · doi:10.2307/3318721
[10] Mosler, K. (2002). Multivariate Dispersion, Central Regions, and Depth: The Lift Zonoid Approach . Lecture Notes in Statist. 165 . Springer, New York. · Zbl 1027.62033
[11] Oja, H. (1983). Descriptive statistics for multivariate distributions. Statist. Probab. Lett. 1 327–332. · Zbl 0517.62051 · doi:10.1016/0167-7152(83)90054-8
[12] Oja, H. (1999). Affine invariant multivariate sign and rank tests and corresponding estimates: A review. Scand. J. Statist. 26 319–343. · Zbl 0938.62063 · doi:10.1111/1467-9469.00152
[13] Ollila, E., Hettmansperger, T. P. and Oja, H. (2002). Affine equivariant multivariate sign methods. Unpublished manuscript. · Zbl 1090.62052
[14] Ollila, E., Oja, H. and Croux, C. (2002). The affine equivariant sign covariance matrix: Asymptotic behavior and efficiency. Unpublished manuscript. · Zbl 1044.62063
[15] Ollila, E., Oja, H. and Hettmansperger, T. P. (2002). Estimates of regression coefficients based on the sign covariance matrix. J. R. Stat. Soc. Ser. B Stat. Methodol. 64 447–466. JSTOR: · Zbl 1090.62052 · doi:10.1111/1467-9868.00344 · links.jstor.org
[16] Peters, D. and Randles, R. H. (1990). A multivariate signed-rank test for the one-sample location problem. J. Amer. Statist. Assoc. 85 552–557. JSTOR: · Zbl 0709.62051 · doi:10.2307/2289797 · links.jstor.org
[17] Randles, R. H. (1989). A distribution-free multivariate sign test based on interdirections. J. Amer. Statist. Assoc. 84 1045–1050. JSTOR: · Zbl 0702.62039 · doi:10.2307/2290081 · links.jstor.org
[18] Visuri, S., Koivunen, V. and Oja, H. (2000). Sign and rank covariance matrices. J. Statist. Plann. Inference 91 557–575. · Zbl 0965.62049 · doi:10.1016/S0378-3758(00)00199-3
[19] Visuri, S., Ollila, E., Koivunen, V., Möttönen, J. and Oja, H. (2003). Affine equivariant multivariate rank methods. J. Statist. Plann. Inference 114 161–185. · Zbl 1011.62053 · doi:10.1016/S0378-3758(02)00469-X
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.