1. Forum
  2. Usenet
  3. SCI.OP-RESEARCH

  • Becareful of USL for R

    From Ramine@21:1/5 to All on Thu Apr 14 18:46:35 2016
    Hello,


    Please look at this R package from the Dr. Gunther the author of USL:

    https://cran.r-project.org/web/packages/usl/


    Becareful of this USL package , because the nonlinear regressions
    that are supported by R and that is using this USL package is
    not efficient, because you can not set the bounds on the the regression parameters between 0.0 and 1.0 , this is required for USL, so the
    nonlinear regression of this R package is not efficient..

    so i have thought more about this problem , and i am currently
    implementing the Simplex method as a minimization method for the
    the nonlinear regression and this will allow me to set the bounds
    of the regression parameters, and this will work much more efficiently
    and better than the R package , this is all about mathematics and i am currently implementing it in my USL programs, i have just tested it and
    it is working much more efficiently than the R package for USL.

    It will take me one day to implement this nonlinear regression using
    the Simplex method as minimization method.

    So i will provide you with the best USL programs around.



    Thank you,
    Amine Moulay Ramdane.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Ramine@21:1/5 to All on Thu Apr 14 18:59:03 2016
    Hello,


    Sorry i have made a mistake:

    Because look at this:

    https://cran.r-project.org/web/packages/nlmrt/nlmrt.pdf


    As you have noticed you can set the the bounds of the parameters
    with the nlxb R function, here it is:

    nlxb(formula, start, trace=FALSE, data, lower=-Inf, upper=Inf,
    masked=NULL, control, ...)


    So the R pacakge for USL is efficient.


    Thank you,
    Amine Moulay Ramdane.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)