logpost.Gauss.compute(object, theta, prior, rcmin)
"preGauss"
.
paramIter
component of a
missmodel
object (see
)
"priorGauss"
giving the hyperparameters
of the prior distribution. Routine
priorGauss
is used to create
the class
"priorGauss"
object. Alternatively, the character
strings "ml" (for no prior, i.e., maximum likelihood estimation),
"noninformative" (for a noninformative prior), or "ridge" (for the
default ridge prior) may be used. Pattern matching means that only
the first character in the string is required. See
for details.
sqrt(.Machine$double.eps)
is acceptable.
This routine is not normally called directly by users.
Schafer, J. L. (1997), Analysis of Incomplete Multivariate Data, Chapman & Hall, London.