init.rmultiregfp {bayesm} | R Documentation |
init.rmultiregfp
initializes variables which can be pre-computed
for draws from the posterior of a multivariate regression model.
init.rmultiregfp
should be called prior to use of rmultiregfp
init.rmultiregfp(X, A, Bbar, nu, V)
X |
Design matrix |
A |
Prior Precision matrix (m x k) |
Bbar |
Prior mean matrix (m x k) |
nu |
degrees of freedom parmeter for Sigma prior |
V |
location parameter for Sigma prior |
model: Y = XB + U. u_i ~ N(0,Σ). u_i is the ith row of U. Y is n x m. X is n x k. B is k x m.
priors: vec(B) ~ N(vec(Bbar,Σ (x) A^{-1})
Σ ~ IW(nu,V).
A list containing ...
IR |
Inverse of Cholesky Root of (X'X + A) |
RA |
Cholesky root of A |
RABbar |
RA %*% Bbar |
nu |
d.f. parm for IWishart prior |
V |
location matrix for IWishart prior |
Peter Rossi, Graduate School of Business, University of Chicago, Peter.Rossi@ChicagoGsb.edu.
For further discussion, see Bayesian Statistics and Marketing
by Rossi, Allenby and McCulloch, Chapter 2.
http://gsbwww.uchicago.edu/fac/peter.rossi/research/bsm.html