rmvnorm(n , p), integers n > 0, p > 1 rmvnorm(n, sigma), REAL positive definite symmetric matrix sigma, nrows(sigma) > 1 rmvnorm(n, p or sigma, mu), REAL row or column vector mu with length(mu) = p or ncols(sigma) |

rmvnorm(n, p), where n > 0 and p > 0 are integer scalars, returns an n by p matrix whose rows are a random sample from the standard multivariate normal distribution MVN(0, I_p), where I_p is the p by p identity matrix. rmvnorm(n, p, mu) where mu is a REAL row or column vector of length p does the same, except the rows of the result are MVN(mu, I_p). rmvnorm(n, sigma [,mu]), where sigma is a positive definite p by p symmetric matrix with p > 1, does the same, except the rows of the result are MVN(0, sigma) or MVN(mu, sigma). When p = 1, use mu + sd*rnorm(n) or mu + sd*rmvnorm(n,1) to generate a normal sample with mean mu and standard deviation sd. See also rnorm().

Gary Oehlert 2003-01-15