Extract and label parameters from the
paramIter component of a
missmodel
object.
Function
paramIter is generic.
Methods currently exist for
missmodel objects,
and for the following specific classes of
paramIter components
of
missmodel objects:
cgm,
Gauss and
Loglin.
USAGE:
paramIter(object, ...)
paramIter.missmodel(object, ...)
paramIter.default(object, save = dim(object)[1])
paramIter.cgm(object, save = dim(object)[1], expand = F)
paramIter.Gauss(object, save = dim(object)[1], expand = F)
paramIter.Loglin(object, save = dim(object)[1])
REQUIRED ARGUMENTS:
object
Either a
missmodel object, or a
paramIter component of a
missmodel
object.
OPTIONAL ARGUMENTS:
...
additional arguments to pass to methods for specific classes of
paramIter components.
save
a vector giving the rows (iterations) in the parameter matrix to be extracted.
The default is to extract the last row.
expand
logical value;
if TRUE, then the parameters are expanded into a more readily understood form.
VALUE:
the requested parameters.
For a
Loglin model,
the parameters are organized as an array, where each "row" (the first
dimension) is for an iteration and remaining dimensions form
a table of estimated cell probabilities for that iteration.
For other models,
if
expand=FALSE,
the value is a matrix
with class
"Gauss" or
"cgm"
in which each row contains parameter estimates from one iteration.
If
expand=TRUE the value is a list of
lists, each inner list containing parameters for one iteration.
For a
Gauss model each inner
list contains a
vector of means and a variance-covariance matrix.
For a
cgm model each inner
list contains a
matrix of means (one row for each continuous variable,
and one column for each cell in the contingency table),
a variance-covariance matrix, and a vector of cell
probabilities.
If only a single iteration is requested, instead of a list of a single
list the single "inner" list is returned; this is subject to change.