Multivariate reciprocal inverse Gaussian distributions from the Sabot-Tarrès-Zeng integral (Q2293388)

The authors define a generalization of the integral \( STZ_n \), called \( MRIG_n \). Thus, forming a new set in \( \mathbb{R}^n \) called \( MRIG_n \) family that is stable by marginalization by conditioning. The \( MRIG_n \) distributions have some properties that are multivariate versions of properties known under the univariate \( RIG \) law. The authors present \( MRIG_n \) integral and its some forms and the Laplacian of \( W \), where \( W = (w_{ij})_{1 \le i, j \le n} \) is the symmetric matrix such that \( w_{ij} = 0 \) for all \( i = 1, \dots, n \), \( w_{ij} \ge 0 \) for \( i \neq j \) and parameterizations of \( C_W \) for \( (0, \infty)^n \) and \( \mathbb{R}^n \). Examples are made that consider some graphs associated with \(W\). A study of the \( MRIG_n \) distributions as a natural exponential family, the marginals, conditional distributions, and a convolution property of the \( MRIG_n \) laws. Another generalization of the Sabot-Tarrés-Zeng integral [\textit{C. Sabot} and \textit{P. Tarrès}, J. Eur. Math. Soc. (JEMS) 17, No. 9, 2353--2378 (2015; Zbl 1331.60185)] is also presented. Finally, the continuity of the density of the laws of \( MRIG_n \) is discussed.