Rev Language Reference
dnMultivariateNormal - Multivariate Normal Distribution
The multivariate normal distribution has the probability density:
f(x) = det(2 pi Sigma)^(-1/2) e^{-(1/2) (x-mu)' Sigma^-1 (x-mu)}
where mu is a vector of mean values and Sigma is a covariance matrix. Note, this distribution may also be parameterized in terms of the precision matrix, Sigma^-1.
Usage
dnMultivariateNormal(Real[] mean, MatrixRealSymmetric covariance, MatrixRealSymmetric precision, RealPos scale)
Arguments
| mean : | Real[] (pass by const reference) |
| The vector of mean values. | |
| covariance : | MatrixRealSymmetric (pass by const reference) |
| The variance-covariance matrix. | |
| Default : NULL | |
| precision : | MatrixRealSymmetric (pass by const reference) |
| The precision matrix. | |
| Default : NULL | |
| scale : | RealPos (pass by const reference) |
| The scaling factor of the variance matrix. | |
| Default : 1 |
Domain Type
Example
dim = 4
df = 100
kappa <- 2
Sigma ~ dnWishart(df, kappa, dim)
for (i in 1:dim) { mu[i] ~ dnUnif(-1, 1) }
x ~ dnMultivariateNormal( mean=mu, covariance=Sigma )
mv[1] = mvCorrelationMatrixElementSwap(Sigma)
mv[2] = mvCorrelationMatrixRandomWalk(Sigma)
mv[3] = mvCorrelationMatrixSingleElementBeta(Sigma)
mv[4] = mvCorrelationMatrixSpecificElementBeta(Sigma)
mv[5] = mvCorrelationMatrixUpdate(Sigma)
mv[6] = mvVectorSlide(x)
Edit this page