I'm usually fitting curve on Matlab using "fminsearch" function, which is a really useful and powerful function.
As I'm currently more using R than Matlab... I wonder if that kind of function or script exist on R ?
That would be perfect if you can provide an example conduct on Matlab and on R.
Regards
see the pracma package and the note at
http://www.inside-r.org/packages/cran/pracma/docs/fminsearch
but also package neldermead may have one implementation.
see the pracma package and the note at
http://www.inside-r.org/packages/cran/pracma/docs/fminsearch
but also package neldermead may have one implementation.
If you are looking for some sort of clone to MATLAB's fminsearch, pracma package may be a good starting point as the answer above pointed out. If you are working in optimization problems in general (i do not know what specific kind of optimization problem you have at hand) you can take a look at the optim() function first from the default stats package. The help page says:
"General-purpose optimization based on Nelder-Mead, quasi-Newton
and conjugate-gradient algorithms. It includes an option for
box-constrained optimization and simulated annealing."
There are plenty of examples on the web, just search for "R optim() example".
In any case, the procedure is rather straightforward:
(1) define the function to be optimized
(2) set initial values for the parameters
(3) select the method (e.g. Nelder-Mead)
(4) run optim(par, function, method)
Hope this helps. Cheers!
Look also here:
http://people.hss.caltech.edu/~foley/rtut/week5/optimization_examples.R
Thank you all for your answers...
I'll have a close look at each option.
Thanks again !
Do look at optimx package by John Nash. I found it really helpful.
http://cran.r-project.org/web/packages/optimx/index.html
- Badges
- Science topic