Density expansions for correlations and eigenvalues of the covariance matrix

  • Tõnu Kollo University of Tartu
  • Anne Selart University of Tartu
Keywords: skew normal distribution, multivariate density expansion, multivariate cumulants, multivariate Hermite polynomials

Abstract

In the paper explicit approximation formulae for the sample correlation coefficient and eigenvalues of the sample covariance matrix are found. The approximations are based on univariate and multivariate normal and skew normal distributions. A simulation experiment has been carried out where the empirical distributions are compared with different approximations.

Downloads

Download data is not yet available.

Author Biographies

Tõnu Kollo, University of Tartu

Insitute of Mathematical Statistics

Anne Selart, University of Tartu

Insitute of Mathematical Statistics

Published
2004-12-31
Section
Articles