1 January 2015 in International Journal of Mathematics and Mathematical Sciences
Hypergeometric functions of matrix arguments occur frequently in multivariate statistical analysis. In this paper, we define and study extended forms of Gauss and confluent hypergeometric functions of matrix arguments and show that they occur naturally in statistical distribution theory.
15 April 2008 in Statistics
Two new families of matrix variate distributions are introduced. They are based on matrix normal distribution and yet can be used to model data involving skewness. The properties of the two families are investigated. Among others, the marginals, conditionals, stochastic representation, linear and quadratic forms are studied.
9 June 2013 in Journal of Statistical Theory and Practice
Like some other multivariate skew normal distributions, the unified skew normal (SUN) distribution preserves important properties of the normal distribution. In this article, we show that for a random vector with a unified multivariate skew normal distribution all column (row) full rank linear transformations are in the same family of distributions. Using this property, we provide a characterization for the SUN distribution. In addition, we show that the joint distribution of the independent SUN random vectors is again a SUN distributed random vector. With this property and closure under the linear transformation, we finally show the closure of sums of independent SUN random vectors, and as expected it belongs to the same family.
9 December 2003 in Biometrical Journal
1 April 2011 in Statistics
In this paper, we suggest two estimators of the precision matrix of the multivariate elliptically contoured stable distribution. The properties of these two estimators have been studied theoretically as well numerically. Conditions are given when one estimator dominates the other. Furthermore, trace of the precision matrix has been also estimated.
14 June 2012 in National Academy Science Letters