Extracts the fitted mean or the fitted linear predictor values for
gee
objects.
USAGE:
fitted.gee(object, type="mean")
This is a method for the function
fitted() for objects inheriting
from class
"gee".
REQUIRED ARGUMENTS:
object
an object inheriting from class
"gee" representing a fitted model.
OPTIONAL ARGUMENTS:
type
a character string indicating the desired type of fitted values,
with choices
"mean" for fitted mean (default)
and
"link" for fitted linear predictor.
VALUE:
The
mean fitted values are the expected mean
values of the observations under the given GEE model.
The
link fitted values are the
calculation of the linear predictor.
The relationship between the link values and the mean values is given by
mean = h(link), where h is the link function of the model.
SEE ALSO:
,
,
,
EXAMPLES:
Seizure.Subject <- recordDesign(cluster = "Subject", data = Seizure)
gee.out <- gee(y~Time+group, cluster=cbind(clusterID,recordID),
variance="glm.scale", correlation="exchangeable",
family="poisson", link="log", data=Seizure.Subject)
# produce a plot of fitted versus model residuals
plot(fitted(gee.out, type="mean"), residuals(gee.out,type="model"),
ylab="model residuals")