[R] Please Help for Augmented Prediction Plot

Dieter Menne dieter.menne at menne-biomed.de
Wed Jan 28 10:27:08 CET 2009


Chun-Hao Tu <tch28 <at> hotmail.com> writes:

> Hi R users,I have a question about augmented prediction plot (?augPred). 
> The covariate of my data set is  c(0, 0.01, 0.1, 1, 10, 100, 1000) and I 
> have fitted a nonlinear mixed effects model. I use
> plot(augPred(out.nlme....)) to get the augmented prediction plot. However,
> because the scale of the  covariate is too large thus I am not able to see 
> the detail difference at c(0,0.01, 0.1, and 1). Could anyone
> tell me how to enlarge the plot at that range "c(0,0.01, 0.1, and 1)" ?  


Very, very likely, if you have such a large range (frequently a drug dose), 
you should think of doing a (shifted log?) transform of your data initially. 
Try to do a residual plot plot(result(nlme)) might work first to check for 
this.  Maybe even plotting (0.001,0.01,...) would be more useful.

Otherwise, you could always use pred() manually and do a trellis plot
with some (log+x) transformed data. 

Dieter




More information about the R-help mailing list