QIC for Generalized Estimating Equations
Computes the quasi-likelihood under the independence model criterion (QIC) for one or more objects of the class glmgee.
QIC(..., k = 2, u = FALSE)
... |
one or several objects of the class glmgee which are obtained from the fit of generalized estimating equations. |
k |
an (optional) non-negative value giving the magnitude of the penalty. By default, |
u |
an (optional) logical switch indicating if QIC should be replaced by QICu. By default, |
A matrix with the values of -2*quasi-likelihood, the number of degrees of freedom, and the value of QIC (or QICu if u=TRUE) for each glmgee object provided in the input.
Pan W. (2001) Akaike's information criterion in generalized estimating equations, Biometrics 57, 120-125.
Hin L-Y, Carey V.J., Wang Y-G (2007) Criteria for Working–Correlation–Structure Selection in GEE: Assessment via Simulation. The American Statistician 61, 360–364.
## Example 1
mod <- size ~ poly(days,4) + treat
fit1 <- glmgee(mod, id=tree, family=Gamma("log"), data=spruce, corstr="AR-1")
fit2 <- update(fit1, corstr="Exchangeable")
fit3 <- update(fit1, corstr="Stationary-M-dependent(2)")
fit4 <- update(fit1, corstr="Independence")
QIC(fit1, fit2, fit3, fit4)
## Example 2
mod <- dep ~ visit + group
fit1 <- glmgee(mod, id=subj, family=gaussian, corstr="Exchangeable", data=depression)
fit2 <- update(fit1, corstr="Exchangeable")
fit3 <- update(fit1, corstr="Non-Stationary-M-dependent(2)")
fit4 <- update(fit1, corstr="Independence")
QIC(fit1, fit2, fit3, fit4)
## Example 3
mod <- depressd ~ visit + group
fit1 <- glmgee(mod, id=subj, family=binomial, corstr="Exchangeable", data=depression)
fit2 <- update(fit1, corstr="Exchangeable")
fit3 <- update(fit1, corstr="Stationary-M-dependent(2)")
fit4 <- update(fit1, corstr="Independence")
QIC(fit1, fit2, fit3, fit4)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.