Random beta matrix and the distribution of its determinant (Q1822152)

Let A and I-A be random positive definite Hermitian matrices. Then A is said to have a complex multi-variate beta distribution if the density function of A is a normalizing factor times \(| A|^{\alpha - p}| I-A|^{\beta -p}\) where the real parts of \(\alpha\) and \(\beta\) exceed the dimension p of A and \(| A|\) is the determinant of A. The authors study the density function of \(| A|\). They use the Mellin transform and a particular asymptotic expansion of the ratio of gamma functions written in terms of generalized Bernoulli polynomials.