Cumulants of multiinformation density in the case of a multivariate normal distribution

08/19/2019

∙

by   Guillaume Marrelec, et al.

∙

0

∙

share

We consider a generalization of information density to a partitioning into N ≥ 2 subvectors. We calculate its cumulant-generating function and its cumulants, showing that these quantities are only a function of all the regression coefficients associated with the partitioning.