Usage
ddpPost(y, X, Z, subject, dosemat, numt, typet, Omega, omegaplus, n.iter,
n.burn, n.thin, shapealph, ratebeta, M.init)
Arguments
y
An N x 1 response (of subject-measure
cases)
X
Fixed effects design matrix
Z
Random effects design matrix. Assumed grouped
by subjects
subject
An N x 1 set of subject
identifiers
dosemat
An P x T Anova or Multiple
Membership design matrix linking treatment dosages to
subjects where T is the total number dosages
across all treatments + 1 for an intercept. This
formulation assumes there is a hold-out dose for
numt
A numeric vector of length equal to the
number of treatments that contains the number of dosages
for each treatment.
typet
A numeric vector of length equal to the
number of treatments that contains the base distribution
for each treatment. 1 = "car", 2 = "mvn",
3 = "ind"
Omega
A list object of length equal to the number
of treatments with "car" selected for base
distribution. Each entry is an numt[m] x numt[m]
numeric CAR adjacency matrix for the dosages of treatment
m.
omegaplus
A list object of length equal to the
number of treatments under "car" containing
numeric vectors that are rowSums of the corresponding
matrix element in Omega.
n.iter
The number of MCMC iterations
n.burn
The number of MCMC burn-in iterations to
discard
n.thin
The step increment of MCMC samples to
return
shapealph
The shape parameter for the $\Gamma$
prior on the DP concentration parameter.
ratebeta
The rate parameter for the $\Gamma$
prior on the DP concentration parameter.
M.init
Initial MCMC chain scalar value for number
of by-subject clusters. If excluded defaults to
length(unique(subject)).