Is there a sensible parallel to R^2 or similar model fit statistics for nonlinear regression with nls()
in R
. Is there something implemented in a package? nlstools
package does not offer such statistic but plenty of useful diagnnostics etc.
Asked
Active
Viewed 1,085 times
0

striatum
- 1,428
- 3
- 14
- 31
-
When opening the library I noticed `IMPORTANT NOTICE: Most nonlinear regression models and data set examples related to predictive microbiolgy have been moved to the package 'nlsMicrobio'`--maybe there's a hint? – jay.sf Feb 25 '18 at 17:55
-
1Take a look at https://stats.stackexchange.com/questions/148469/how-should-one-assess-model-fit-for-non-linear-regression – mikeck Feb 25 '18 at 17:59
-
A graphical assessment is often adequate. Assuming `x` is sorted `fm <- nls(y ~ f(x, a, b), start = list(a = 1, b = 1)); plot(y ~ x)); lines(x, fitted(fm))` – G. Grothendieck Feb 25 '18 at 18:25