Multivariate Skew-Normal Generalized Hyperbolic distribution and its properties

JOURNAL OF MULTIVARIATE ANALYSIS

卷:128

Article

Vilca, Filidor¹  Balakrishnan, N.^2,3  Zeller, Camila Borelli⁴ 

[1] Univ Estadual Campinas, Dept Estat, BR-13081970 Sao Paulo, Brazil

[2] McMaster Univ, Dept Math & Stat, Hamilton, ON, Canada

[3] King Abdulaziz Univ, Dept Stat, Jeddah 21413, Saudi Arabia

[4] Univ Fed Juiz de Fora, Dept Estat, Juiz De Fora, MG, Brazil

关键词: Generalized inverse Gaussian distribution;    Skew-normal distribution;    Heavy-tailed distributions;    Skewness and kurtosis;    Normal inverse Gaussian distribution;    Skew-Normal Generalized Hyperbolic distribution;    Mixtures;   

DOI  :  10.1016/j.jmva.2014.03.002

来源: Elsevier

【 摘 要 】

The Generalized Inverse Gaussian (GIG) distribution has found many interesting applications: see Jorgensen [24]. This rich family includes some well-known distributions, such as the inverse Gaussian, gamma and exponential, as special cases. These distributions have been used as the mixing density for building some heavy-tailed multivariate distributions including the normal inverse Gaussian, Student-t and Laplace distributions. In this paper, we use the GIG distribution in the context of the scale-mixture of skew-normal distributions, deriving a new family of distributions called Skew-Normal Generalized Hyperbolic distributions. This new flexible family of distributions possesses skewness with heavy-tails, and generalizes the symmetric normal inverse Gaussian and symmetric generalized hyperbolic distributions. (C) 2014 Elsevier Inc. All rights reserved.

【 授权许可】

Free