Percentage Points of the Multivariate t Distribution

Abstract

The known methods for computing percentage points of multivariate t distributions are reviewed. We believe that this review will serve as an important reference and encourage further research activities in the area.

Tests of heteroscedasticity and correlation in multivariate t regression models with AR and ARMA errors

“…From Liu and Rubin [23], the multivariate t model (1) can be treated as a hierarchical structure of normal and gamma distributions, i.e. given the weight τ , Y has the multivariate normal distribution, and the weight τ is Gamma distributed: [23] presented ML estimation of the multivariate t distribution using EM and its extensions, ECM and ECME; Fernandez and Steel [13] considered likelihood-based inference from multivariate regression models with independent Student-t errors with unknown degrees of freedom; Nadarajah and Kotz [25] reviewed the known methods for computing percentage points of multivariate t distributions. Similar to the normal models, the assumption of homogenous variance is not necessarily appropriate in some situations and needs to be verified.…”