dataDepPrior.preGauss(object, scale, df = 1)
"preGauss"
.
df
degrees of freedom. Any scale
matrix of the correct dimensionality for your problem is
acceptable. Alternatively, if a scalar value is input for
scale
, then
the final scale matrix is taken to be the diagonal matrix containing the
scalar value times the variances of the
nonmissing data. (i.e. For each variable, missing values are deleted,
and variances calculated on just the observed data.)
The usual biased maximum likelihood estimates are computed,
see routine
colVars
.
"priorGauss"
containing the expanded scale matrix
and other hyperparameters. The components in this class are described in
the help file for
.
Schafer, J. L. (1997), Analysis of Incomplete Multivariate Data, Chapman & Hall, London.
dataDepPrior.preGauss(object = preGauss(marijuana), scale = 1.4, df = 2)