[R-sig-ME] help with pdDiag and nlme

Anne Hoen anniehoen at gmail.com
Thu Feb 6 23:17:38 CET 2014


Hi,
I am trying to fit a non-linear mixed effects model using a 3-parameter logistic function. 
My data set can be made using this:

plate<-c(rep("A", 6), rep("B", 6), rep("C", 6), rep("D", 6))
preparation<-rep(c(rep("reference", 3), rep("test", 3)), 4)
concentration<-rep(c(1, 2, 3), 8)
OD<-runif(24, 0, 1)
data<-data.frame(plate, preparation, concentration, OD)
data<-groupedData(OD~concentration | plate, data=data)

I want to fit a model like this:

nlme(OD~SSlogis(concentration, Asym, xmid, scal), 
     fixed=preparation~1, 
     random=pdDiag(plate~1),
     data=data)

But I get the error: Error in nlme.formula(OD ~ SSlogis(concentration, Asym, xmid, scal), fixed = preparation ~  : 
  subscript out of bounds

I've determined that (I think) the error comes from the random effect specification, because when I run: pdDiag(data$plate~1), I get: Uninitialized positive definite matrix structure of class pdDiag, which is not what I expected. I am not sure what is wrong. If anyone can help me, I would appreciate it so much!!

Thanks
Annie


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