MLIC and MLICC for Weighted GEE
Calculate the MLIC (missing longitudinal information criterion) for selection of mean model, and the MLICC (missing longitudinal information correlation criterion) for selection of working correlation structure, based on the expected quadratic loss and the WGEE.
MLIC.gee(object,object_full)
object |
a fitted model object of class |
object_full |
a fitted model object of class |
Return a data frame of MLIC, MLICC and Wquad_loss.
MLIC and MLICC model selection criterion for longitudinal data criterion with dropouts or monotone missingness under the assumption of MAR.
Cong Xu, Zheng Li and Ming Wang
Robins, J.M., Rotnitzky, A. and Zhao, L.P., 1995. Analysis of semiparametric regression models for repeated outcomes in the presence of missing data. Journal of the American Statistical Association, 90(429), pp.106-121.
Shen, C.W. and Chen, Y.H., 2012. Model selection for generalized estimating equations accommodating dropout missingness. Biometrics, 68(4), pp.1046-1054.
Shen, C.W. and Chen, Y.H., 2013. Model selection of generalized estimating equations with multiply imputed longitudinal data. Biometrical Journal, 55(6), pp.899-911.
data(imps)
fit1 <- wgee (Y ~ Drug+Sex+Time,data=imps,id=imps$ID,family="binomial",
corstr="exchangeable",scale=NULL,mismodel= R ~ Drug+Time)
fit_f <- wgee (Y ~ Drug+Sex+Time+Time*Sex+Time*Drug,data=imps,id=imps$ID, family="binomial",
corstr="exchangeable",scale=NULL,mismodel= R ~ Drug+Time)
###not run#####
##MLIC.gee(fit1,fit_f)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.