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)