Hello, I am using SSfol in nlme to fit some data for the change of N concentration (N) in plant tissue over time (gdd). The model works nicely for 2 out of 3 treatments, so I would really like to use it, but it consistently has a bad fit for my third treatment. I am pasting the figure for the third treatment only. I feel that I have my fixed and random effects properly identified, but have also tried many combinations to see if I can improve the fit.
Are there any other ideas of what I can do to capture the highest point of N with the model? http://r.789695.n4.nabble.com/file/n4615613/Predicted_N_2.png Here is my code for the figure followed by a link to the dataset. library(nlme) cna<-read.table("aboveground C and N, dates removed, zeros added.txt", header=TRUE) ###Aboveground 2011, treatment PF only n11a<-subset(cna, cna$year == "2011" & cna$ground == "Above" & cna$trt == "PF") n11a$plotF <- as.factor(n11a$plot) n11aG<-groupedData(N ~ gdd | plotF, data=n11a) fit.dose<- nlsList(N ~ SSfol(dose, gdd, lke, lka, lCl), data = n11aG) plot(intervals(fit.dose), layout=c(3,1)) fit.nlme.11a<- nlme(fit.dose, random=pdDiag(lCl ~ 1)) fit.nlme2.11a<-update(fit.nlme.11a, random = list(lCl + lka ~ 1)) png("thermaltimefigs/Predicted N 1.png") plot(augPred(fit.nlme.11a), level = 0:1) dev.off() png("thermaltimefigs/Predicted N 2.png") plot(augPred(fit.nlme2.11a), level = 0:1) dev.off() data at: http://dl.dropbox.com/u/21080842/aboveground%20C%20and%20N%2C%20dates%20removed%2C%20zeros%20added.txt Thanks, Ranae -- View this message in context: http://r.789695.n4.nabble.com/SSfol-in-nlme-tp4615613.html Sent from the R help mailing list archive at Nabble.com. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.