Inference for the Covariance and Correlation Matrices of Multivariate Sample Using Wishart Distribution

The purpose of this article is to introdused the inference techniques for the mean vector mu the correlation matrix rho and the covariance matrix Sigma of the multivariate normal sample and to apply these techniques using the software package STATISTICA 13.0 (StatSoft Inc, USA). The sample contains 50 weekly return observations (in percent) on each of ten stock portfolios constructed from stocks on the Toronto Stock Exchanges. Since the data are obtained as a random sample of multivariate normal distribution the Wishart distribution can be used to make inference about covariance matrix.