Thanks for you comment. There is certainly some "Monte Carlo sampling" involved in mvtnorm but why derivatives could not be computed? In theory, the derivatives exist (eg. bivariate probit). Moreover, when used with optim, there are some numerical derivatives computed... does it mean that mvtnorm cannot be used in an optimisation problem? I think it hard to believe.
One possibility would be to use the analytical derivatives and then a do-it-yourself integration but i was looking for something a bit more comprehensive. The mvtnorm package uses a specific way to compute pmvnorm and I'm far to do a good enough job so that derivatives can compare with what mvtnorm can do. Stef ______________________________________________ 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.
