Does anybody know of a function that implements the derivative (gradient) of
the multivariate normal density with respect to the *parameters*?
It’s easy enough to implement myself, but I’d like to avoid reinventing the
wheel (with some bugs) if possible. Here’s a simple example of the result
I’d like, using numerical differentiation:
library(mvtnorm)
library(numDeriv)
f=function(pars, xx, yy)
{
mu=pars[1:2]
sig1=pars[3]
sig2=pars[4]
rho=pars[5]
sig=matrix(c(sig1^2,rho*sig1*sig2,rho*sig1*sig2,sig2^2),2)
dmvnorm(cbind(x,y),mu,sig)
}
mu1=1
mu2=2
sig1=3
sig2=4
rho=.5
x=2 # or a x vector
y=3 # or a y vector
jacobian(f,c(mu1,mu2,sig1,sig2,rho),xx=x,yy=y)
# (Can replace ‘jacobian’ with ‘grad’ if x and y have length 1.)
--
Karl Ove Hufthammer
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
