Hello everyone,
I'm trying to test the accurracy of R on the Eckerle4 dataset from NIST and
I don't understand how the control option of the nls function works.
I tought nls(...) was equivalent to nls(...control=nls.control()) i.e
nls.control() was the default value of control, but here is the error I get :
> n2=nls(V1~(b1/b2) *
> exp(-0.5*((V2-b3)/b2)^2),data=ecker,start=list(b1=1.5,b2=5,b3=450,control=nls.control()))
Error in nlsModel(formula, mf, start) : singular gradient matrix at initial
parameter estimates
while I get no error without setting the control option with the same other
parameters.
I see that R didn't manage to solve the Eckerle4 regression problem from start
one while Splus can do it with the nlregb option.
Is there something equivalent for R now?
Otherwise, I found that R 2.0.1 was performing better than SAS 9.1 on the NIST
Datasets in general.
Best regards,
Anthony Landrevie
---------------------------------
[[alternative HTML version deleted]]
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
