Skew-normal distribution is an extension of the normal distribution where the symmetry of the normal distribution is distorted with an extra parameter called the shape parameter. The scalar version of the skew-normal distribution was introduced already in works by Azzalini ( 1985 Azzalini, A. (1985). A class of distributions which includes the normal ones. Scand. J. Stat. 12:171–178. [Web of Science ®] [Google Scholar] ) and by Henze ( 1986 Henze, N. (1986). A probabilistic representation of the ’skew-normal’ distribution. Scand. J. Stat. 13:271–275. [Web of Science ®] [Google Scholar] ). The multivariate skew-normal distribution has received considerable attention over the last years, but unfortunately there is no unique straightforward generalization from the scalar case. Therefore, various families of skew-symmetric distributions with different properties have been proposed and studied. In our work we refer to the “classical” multivariate skew-normal distribution introduced by Azzalini and Dalla Valle ( 1996 Azzalini, A. , Dalla Valle, A. (1996). The multivariate skew-normal distribution. Biometrika 83(4):715–726. [Crossref], [Web of Science ®] [Google Scholar] ). It must be noted that even the Azzalini’s skew-normal distribution can be parametrized in many different ways starting from the initial -parametrization to the currently prevalent -parametrization. This motivated us to search for viable alternatives and compare the overall behavior of different parametrizations, the key problems addressed in our article.
Symmetric elliptical distributions have been intensively used in data modeling and robustness studies. The area of applications was considerably widened after transforming elliptical distributions into the skew elliptical ones that preserve several good properties of the corresponding symmetric distributions and increase possibilities of data modeling. We consider three-parameter p-variate skew t-distribution where p-vector is the location parameter, is the positive definite scale parameter, p-vector is the skewness or shape parameter, and the number of degrees of freedom is fixed. Special attention is paid to the two-parameter distribution when that is useful for construction of the skew t-copula. Expressions of the parameters are presented through the moments and parameter estimates are found by the method of moments. Asymptotic normality is established for the estimators of and . Convergence to the asymptotic distributions is examined in simulation experiments.
1 November 2009 in Journal of Statistical Planning and Inference