rmixGibbs {bayesm} | R Documentation |
rmixGibbs
makes one draw using the Gibbs Sampler for a mixture of multivariate normals. rmixGibbs
is not designed to be called directly. Instead, use rnmixGibbs
wrapper function.
rmixGibbs(y, Bbar, A, nu, V, a, p, z)
y |
data array where rows are obs |
Bbar |
prior mean for mean vector of each norm comp |
A |
prior precision parameter |
nu |
prior d.f. parm |
V |
prior location matrix for covariance prior |
a |
Dirichlet prior parms |
p |
prior prob of each mixture component |
z |
component identities for each observation – "indicators" |
a list containing:
p |
draw of mixture probabilities |
z |
draw of indicators of each component |
comps |
new draw of normal component parameters |
This routine is a utility routine that does not check the input arguments for proper dimensions and type.
Rob McCulloch (Arizona State University) and Peter Rossi (Anderson School, UCLA), perossichi@gmail.com.
For further discussion, see Chapter 5 Bayesian Statistics and Marketing by Rossi, Allenby, and McCulloch.
http://www.perossi.org/home/bsm-1