Definition and probabilistic properties of skew-distributions
Loading...
Date
2002
Journal Title
Journal ISSN
Volume Title
Publisher
ELSEVIER
Abstract
The univariate, and multivariate skew-normal distributions have a number of intriguing properties. It is shown here that these hold for a general class of distributions, defined in terms of independence conditions on signs and absolute values. For this class, two stochastic representations become equivalent, one using conditioning on the positivity of a random vector and the other employing a vector of absolute values. General methods for computing moments and for obtaining the density function of a general skew-distribution are given. The case of spherical and elliptical distributions is briefly discussed. (C) 2002 Published by Elsevier Science B.V.
Description
Keywords
skew-distributions, skew-normal, spherical distributions, elliptical distributions, stochastic representations