Add covariate in nlme?

[hwouters]

Hi R-experts,

I successfully used you nlme function in R but I’m struggling when including a covariate in the model.
I could implement the same model in sas but need to do the work in R for practical reasons.

I use the following code to group my data:

Parameterg <- groupedData( result ~ time | Batch,
data = Batchdata,
labels = list( x = "Time", y = "analysis")
)

and then uses the nlme function to fit a nonlinear mixed model that includes Process as a fixed covariate:
nlme.model001epr <- nlme(result ~ A0 * exp(- ( exp(A1) * exp(-Ea / (0.0083144*TEMP.K)) * exp(eps)) * time),
data = Parameterg,
fixed=list(A0+Ea~1,A1~Process),
random=eps~1,
start=c(93, 92, 34.5,37),
control=list(msVerbose=TRUE, maxIter = 200),
verbose=TRUE, method="REML",
na.action=na.pass)
this fit give the following error:
Error in MEEM(object, conLin, control$niterEM) :
Singularity in backsolve at level 0, block 1

Note that I’m not having any missing values in my data.

Do you know what the correct implementation is for adding a covariate in a nlme and what might be my problem here?

FYI: the sas-code that worked in my case was like:
proc nlmixed data=stab.BatchdataParameter ;
parms Ea=92 A0=93 A1=34.5 e=10 s2=10 A11=1 A12=1;
k= exp(A1)*exp(-Ea/(.0083144*tempK))*exp(eps1)*exp(A11*(Process='Process 1') + A12*(Process='Process 3'));
m= A0*exp(-k*time);
model result ~ normal(m,e);
random eps1~normal(0,s2) subject=batch;
contrast "Process" A11,A12;
contrast "Process 2 vs 3" A12;
estimate "Process 1" A1+A11;
estimate "Process 3" A1+A12;
estimate "Process 2" A1;
run;

Looking forward to your replies,
Heidi

[hwouters]

Problem solved: my initial values were not chosen correctly :-)