Dear S-Plus users,
I am trying to use the nlme() function to analyze growth
curves by fitting a 2-parameter exponential model.
When I fit separate models by subject using nlsList(), I don’t
have any problem and get estimates for each subject.
However, when I try to use nlme(), I systematically get the
following error message:
Problem in .C("mixed_EM",: Singularity in backsolve,
while calling subroutine mixed_EM
The call is the following:
nlme(tumor.size ~ a*b**Day , fixed = a + b ~ Treatment , data = ""
random = b ~ Treatment, start= c(a=1,b=1))
I am using S-Plus 6.1 on Windows XP SP2.
I have tried other nonlinear functions as well as fitting
the model on a subset of subjects without success.
I found a bug report on Insightful’s web site for a
similar problem but with lme() but nothing concerning nmle().
Does anyone have any idea on this issue?
Thanks in advance
Michel Guillet