Ddivergence: d-divergence

In MCMCglmm: MCMC Generalised Linear Mixed Models

d-divergence

Description

Calculates Ovaskainen's (2008) d-divergence between 2 zero-mean multivariate normal distributions.

Usage

Ddivergence(CA=NULL, CB=NULL, n=10000)

Arguments

CA	Matrix A
CB	Matrix B
n	number of Monte Carlo samples for approximating the integral

Value

d-divergence

Note

In versions of MCMCglmm <2.26 Ovaskainen's (2008) d-divergence was incorrectly calculated.

Author(s)

Jarrod Hadfield [e-mail]

References

Ovaskainen, O. et. al. (2008) Proc. Roy. Soc - B (275) 1635 593-750

Examples

CA<-rIW(diag(2),10, n=1)
CB<-rIW(diag(2),10, n=1)
Ddivergence(CA, CB)

MCMCglmm documentation built on July 9, 2023, 5:24 p.m.