Abstract
The nth order integral over a generalized quadrant of the multivariate normal distribution has first been transformed to a series of integrals over the entire hypervolume for all orders. The series has been derived for all n and is exhibited explicitly for all orders up to n = 6. The fourth-order integral has further been reduced to three first-order integrals, thus making them amenable to machine solution. This method avoids the difficulty of trying to integrate a mixed distribution over an interval when the variables cannot be separated. Three very special cases of the sixth-order integral are exhibited. It can be shown that the sixth-order integral can be reduced to five second-order integrals.

This publication has 0 references indexed in Scilit: