Erdelyi-Kober Fractional Integral Operators from a Statistical Perspective -II

A.M. Mathai,H.J. Haubold
DOI: https://doi.org/10.48550/arXiv.1303.3979
2013-03-16
Abstract:In this article we examine the densities of a product and a ratio of two real positive definite matrix-variate random variables $X_1$ and $X_2$, which are statistically independently distributed, and we consider the density of the product $U_1=X_2^{1\over2}X_1X_2^{1\over2}$ as well as the density of the ratio $U_2=X_2^{1\over2}X_1^{-1}X_2^{1\over2}$. We define matrix-variate Kober fractional integral operators of the first and second kinds from a statistical perspective, making use of the derivation in the predecessor of this paper for the scalar variable case, by deriving the densities of products and ratios where one variable has a matrix-variate type-1 beta density and the other variable has an arbitrary density. Various types of generalizations are considered, by using pathway models, by appending matrix variate hypergeometric series etc. During this process matrix-variate Saigo operator and other operators are also defined and properties studied.
Classical Analysis and ODEs
What problem does this paper attempt to address?