An object returned by the
gee function
representing a model fit using
the generalized estimating equation approach.
GENERATION:
A
"gee" object is created with a call to function
gee
.
METHODS:
The class
"gee" has the methods
print
,
summary,
coef
,
fitted
,
and
residuals.
VALUE:
a list with the following components:
coefficients
the estimated regression parameters.
correlation
the estimated correlation parameters.
scale
the estimated or fixed scale parameters.
var.random
the estimated variance of random-effects parameters in mixed models.
cor.work
the estimated working correlation matrix.
iter
the number of iterations used to obtain the results.
fitted.values
the fitted response.
variance.y
variance of response, scaled by the final value of the scale factor.
variance
a list of variance estimates for the regression
and correlation parameter estimates from both
the model-based (naive) estimator
and the empirical (sandwich) estimator.
The default will print the sandwich estimate.
For independent and non-clustered data, use
the naive estimate, since the sandwich estimate
will be all zeros.
call
the call that produces the object.
family
the family used in the model.
cor.design
the design of correlation matrix in the call, represented as a
corDesign
object, see
corDesign.object
for details.
control
the settings used to control the optimization, see
geeControl
for details.
n.clusters
the number of distinct clusters.
n.obs
the number of total observations in the data.
initial
the initial values of the regression, correlation, and scale parameters
used in the fitting.