Learn R Programming

mosaic (version 0.14)

predict.groupwiseModel: Evaluate a groupwise model given new data

Description

If newdata is not specified, the data originally used for fitting will be used.

Usage

"predict"(object, newdata = object$data, type = c("class", "likelihood", "prob"), level = NULL, ...)

Arguments

object
a groupwise model
newdata
new data from which to compute fitted valeus.
type
one of "class", "likelihood", or "prob"
level
an optional character string specifying the level for which probabilities are to be reported. Defaults to the first class of the potential outputs. Set to ".all" to see probabilities for all levels.
...
additional arguments (currently ignored)

Details

setting the type is needed only for classifiers. "class" will give just the class as output. "likelihood" will give the probability of the observed outcome (in newdata) given the model. "prob" will give the probability of the class named in level