Hi.
In part answer to my own question, I just found this link:
https://stat.ethz.ch/pipermail/r-help/2003-August/037654.html
which explains how one might calculate an approximation to the
variance of a non-linear function and use that to calculate a
confidence/prediction interval. If anyone knows an easier way, then
I'd still be glad to hear from you.
Thanks,
Paul
On 21/06/07, Paul Suckling <paul.suckling at gmail.com>
wrote:> Hello.
>
> Does anyone out there know if there is a function (or functions) that
> will enable me to calculate prediction and confidence intervals from a
> non-linear least squares model? (predict.nls does not do this.)
> Thank you,
>
> Paul
>