Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We investigate the Oja sign covariance matrix (Oja SCM) for estimating partial correlations in multivariate data. The Oja SCM estimates directly a multiple of the precision matrix and is based on the concept of Oja signs, a multivariate generalisation of the univariate sign function, which obey some form of affine equivariance property. Simulations show that the asymptotic distribution gives a good approximation of the exact finite-sample distribution already for samples of moderate size. We find it to outperform the classical sample partial correlation in case of heavy-tailed distributions. The high computational costs are its main disadvantage.

Original language

English

Title of host publication

Compstat 2008: Proceedings in Computational Statistics. Vol. II.

Compstat 2008: Proceedings in Computational Statistics. Vol. II. : Proceedings in Computational Statistics. ed. / Paula Brito. Vol. 2 Physica-Verlag HD, 2008. p. 721-729.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Vogel, Daniel ; Fried, Roland . / Estimating partial correlations using the Oja sign covariance matrix. Compstat 2008: Proceedings in Computational Statistics. Vol. II. : Proceedings in Computational Statistics. editor / Paula Brito. Vol. 2 Physica-Verlag HD, 2008. pp. 721-729

@inproceedings{a951cb400b5547578ad3e931ad634b36,

title = "Estimating partial correlations using the Oja sign covariance matrix",

abstract = "We investigate the Oja sign covariance matrix (Oja SCM) for estimating partial correlations in multivariate data. The Oja SCM estimates directly a multiple of the precision matrix and is based on the concept of Oja signs, a multivariate generalisation of the univariate sign function, which obey some form of affine equivariance property. Simulations show that the asymptotic distribution gives a good approximation of the exact finite-sample distribution already for samples of moderate size. We find it to outperform the classical sample partial correlation in case of heavy-tailed distributions. The high computational costs are its main disadvantage.",

author = "Daniel Vogel and Roland Fried",

year = "2008",

language = "English",

volume = "2",

pages = "721--729",

editor = "Paula Brito",

booktitle = "Compstat 2008: Proceedings in Computational Statistics. Vol. II.",

publisher = "Physica-Verlag HD",

}

TY - GEN

T1 - Estimating partial correlations using the Oja sign covariance matrix

AU - Vogel, Daniel

AU - Fried, Roland

PY - 2008

Y1 - 2008

N2 - We investigate the Oja sign covariance matrix (Oja SCM) for estimating partial correlations in multivariate data. The Oja SCM estimates directly a multiple of the precision matrix and is based on the concept of Oja signs, a multivariate generalisation of the univariate sign function, which obey some form of affine equivariance property. Simulations show that the asymptotic distribution gives a good approximation of the exact finite-sample distribution already for samples of moderate size. We find it to outperform the classical sample partial correlation in case of heavy-tailed distributions. The high computational costs are its main disadvantage.

AB - We investigate the Oja sign covariance matrix (Oja SCM) for estimating partial correlations in multivariate data. The Oja SCM estimates directly a multiple of the precision matrix and is based on the concept of Oja signs, a multivariate generalisation of the univariate sign function, which obey some form of affine equivariance property. Simulations show that the asymptotic distribution gives a good approximation of the exact finite-sample distribution already for samples of moderate size. We find it to outperform the classical sample partial correlation in case of heavy-tailed distributions. The high computational costs are its main disadvantage.

M3 - Conference contribution

VL - 2

SP - 721

EP - 729

BT - Compstat 2008: Proceedings in Computational Statistics. Vol. II.