[R-sig-ME] generalized linear mixed model for qualitative in R

Fan, Ruzong (NIH/NICHD) [E] ruzong.fan at nih.gov
Sun Feb 12 00:58:52 CET 2012


Dear folks,

I wonder if  glmer can we do spline or model correlation for longitudinal data?

In lme, it is possible to do both spline and model correlation. For instance, I wrote a short codes as below: 
#################################################################
fit1E  <- lme( sbp ~ -1 + X.mean,
                random      = list(group = pdIdent(~-1+Z.mean), 
                                      id = ~1),
                correlation = corExp(form = ~ x | group/id), 
                    na.action   = na.exclude,
                method      = "ML",
                subset      = notmissing )
#################################################################
Basically, 'X.mean' models the fixed effect, and `Z.mean' models the spline random effect. The above codes will lead to 3 random variance estimations: one from Z.mean, one from id, and the other from the residual. 

In addition, 'correlation = corExp(form = ~ x | group/id)' models the correlation. 

I don't see there is something like `random = list( ... )' in glmer. So I am not sure how to do spline using glmer. 

Plus, I don't find similar things like `correlation = corExp(form = ~ x | group/id)' in glmer to model correlation?

Thanks. R.F



More information about the R-sig-mixed-models mailing list