mcsm (version 1.0)

rmunorm: Random generator for the multivariate normal distribution

Description

This function produces one random vector distributed from the multivariate normal distribution $N(mu,sig)$.

Usage

rmunorm(mu, sig)

Arguments

mu

Mean $mu$ of the normal distribution

sig

Covariance matrix $sig$ of the normal distribution

Value

This function returns a real vector of the same dimension as $mu$.

Warning

Similar to dmunorm, this function is fragile in that it does not test for

1. the fact that sig is a square matrix,
2. the compatibility of the dimensions of x, mu, sig
3. the symmetry nor the invertibility of the matrix sig

It is therefore prone to fail if those conditions are not satified! If the package bayesm can be installed, rmvnorm is to be prefered to rmunorm.

References

Chapter 8 of EnteR Monte Carlo Statistical Methods

See Also

rnorm,dmunorm,rmvnorm(bayesm)

Examples

test=NULL
for (t in 1:10^4) test=rbind(test,rmunorm(rep(1,2),matrix(c(1,-2,-2,10),ncol=2)))
cor(test[,1],test[,2])*sqrt(10)  # should be close to -2