[R] nonlinear least squares trust region fitting ?

[Spencer Graves]

      1.  "Port" is NOT and R package but something more generally 
available.  I just got 189 hits from Google for "nl2sol Port package", 
some of which should answer your questions about that. 

      2.  Have you considered 'nlminb' and 'optim'?  There is also a 
"sequential quadratic programming" algorithm embedded in the code for 
'garchFit{fSeries}'. 

      3.  If you'd like more help from this listserve, please post 
another question.  When you do so, please provide commented, minimal, 
self-contained, reproducible code, as suggested in the posting guide 
"www.R-project.org/posting-guide.html".  Please also help us understand 
why you don't want Gauss-Newton -- in prose as simple and clear as 
possible.  Doing so will increase your chances of a prompt reply that 
will likely be closer to what you want. 

      Hope this helps. 
      Spencer Graves

Martin Ivanov wrote:
> Hello!
>
> I am running R-2.3.1-i386-1 on Slackware Linux 10.2. I am a former matlab user, moving to R. In matlab, via the cftool, I performed nonlinear curve fitting using the method "nonlinear least squares" with the "Trust-Region" algorithm and not using robust fitting. Is it possible to perform the same analysis in R? I read quite a lot of R documentation, but I could not find an alternative solution. If there is such, please forgive my ignorance (I am a newbie in R) and tell me which function from which package is capable of performing the same analysis. If the same analysis is not possible to carry out in R, I would be grateful if you suggest to me some alternative procedure. I found that the "nls" function performs nonlinear least squares. The problem is that I do not want to implement the Gauss-Newton algorithm. In the worst case I would be contented with the "Levenberg-Marquardt" algorithm, if it is implemented in R. R nls's documentation mentions the "port" package and the ‘nl
>  2sol’ algorithm, but I could not find that package in the CRAN repository, so that I could read and judge whether that algorithm would be appropriate.
>
> Thank you very much in advance. I am looking forward to your answer.
> Regards,
> Martin
>
> -----------------------------------------------------------------
> http://ide.li/ - портал за българите по света. Статии, новини, форуми, снимки, информация.
>
> ______________________________________________
> R-help at stat.math.ethz.ch 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.
>