cgmDesign(object)
"missmodel" with a
paramIter component that
inherits from class
"cgm".
The conditional gaussian model allows the multivariate normal means to
be predicted as a linear model. This routine returns the design
matrix (
X) in the linear model (
mu = X b).
stlouis.da3 <- daCgm(stlouis3[,-1],
margins = ~risk:D1 + risk:D2 + D1:D2,
design = ~D1*D2 + G,
optData = data.frame(G = rep(-1:1, 4)),
control = list(save = 101:5100, trace = T))
cgmDesign(stlouis.da3)