Powell's Metod

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Powell's Metod

Tolga Uzuner
Folks,
Has anyone implemented Powell's Method for minimisation in R ?
Many thanks,
Tolga

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Reply | Threaded
Open this post in threaded view
|

Re: Powell's Metod

Prof Brian Ripley
(Michael) Powell (of Harwell and Cambridge) invented several optimization
methods.  You will need to give us a precise reference.

Powell is associated with earlier versions of the Nelder-Mead, CG and BFGS
methods implemented in optim() but in each case different variants have
best stood the course of time.

On Sun, 15 Jan 2006, Tolga Uzuner wrote:

> Has anyone implemented Powell's Method for minimisation in R ?

--
Brian D. Ripley,                  [hidden email]
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Reply | Threaded
Open this post in threaded view
|

Re: Powell's Metod

Tolga Uzuner
Ah, my apologies. I meant the method outlined in section 10.5 of
Numerical Recipes using Powell's heuristic of discarding the direction
of largest decrease. A nice twist would be if linmin (line
minimization)  were to be implemented using the methods described in
10.1-10.3, but that is just a nice-to-have.

 From what I understand, Nelder-Mead is the method in section 10.4
(downhill simplex), CG is section 10.6, and BFGS is section 10.7.

Admittedly, optim is fantastic most of the time. However, I have a
factor model over hundreds of elements, where I am trying to back out
the factor distribution and it appears that the method in 10.5 might work.

Thanks in advance in any case,
Tolga

Prof Brian Ripley wrote:

> (Michael) Powell (of Harwell and Cambridge) invented several
> optimization methods.  You will need to give us a precise reference.
>
> Powell is associated with earlier versions of the Nelder-Mead, CG and
> BFGS methods implemented in optim() but in each case different
> variants have best stood the course of time.
>
> On Sun, 15 Jan 2006, Tolga Uzuner wrote:
>
>> Has anyone implemented Powell's Method for minimisation in R ?
>
>

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Reply | Threaded
Open this post in threaded view
|

Re: Powell's Metod

Prof Brian Ripley
I am not aware of an R implementation of that method.  However, if you
meet the licensing conditions (which for my copy of NR do not allow me to
use the code on a Unix machine even if I typed it in) you could interface
to the NR C/C++ routines.

On Sun, 15 Jan 2006, Tolga Uzuner wrote:

> Ah, my apologies. I meant the method outlined in section 10.5 of
> Numerical Recipes using Powell's heuristic of discarding the direction
> of largest decrease. A nice twist would be if linmin (line
> minimization)  were to be implemented using the methods described in
> 10.1-10.3, but that is just a nice-to-have.
>
> From what I understand, Nelder-Mead is the method in section 10.4
> (downhill simplex), CG is section 10.6, and BFGS is section 10.7.
>
> Admittedly, optim is fantastic most of the time. However, I have a
> factor model over hundreds of elements, where I am trying to back out
> the factor distribution and it appears that the method in 10.5 might work.
>
> Thanks in advance in any case,
> Tolga
>
> Prof Brian Ripley wrote:
>
>> (Michael) Powell (of Harwell and Cambridge) invented several
>> optimization methods.  You will need to give us a precise reference.
>>
>> Powell is associated with earlier versions of the Nelder-Mead, CG and
>> BFGS methods implemented in optim() but in each case different
>> variants have best stood the course of time.
>>
>> On Sun, 15 Jan 2006, Tolga Uzuner wrote:
>>
>>> Has anyone implemented Powell's Method for minimisation in R ?
>>
>>
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
>

--
Brian D. Ripley,                  [hidden email]
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Reply | Threaded
Open this post in threaded view
|

Re: Powell's Metod

Ravi Varadhan
Tolga,

Have you considered an EM algorithm for your factor analysis problem?  A
reference for this is:

Rubin, D. and Thayer, D. (1982). EM algorithms for ML factor analysis.
Psychometrika, 47(1):69--76.

Hope this is helpful,
Ravi.

> -----Original Message-----
> From: [hidden email] [mailto:r-help-
> [hidden email]] On Behalf Of Prof Brian Ripley
> Sent: Monday, January 16, 2006 9:31 AM
> To: Tolga Uzuner
> Cc: [hidden email]
> Subject: Re: [R] Powell's Metod
>
> I am not aware of an R implementation of that method.  However, if you
> meet the licensing conditions (which for my copy of NR do not allow me to
> use the code on a Unix machine even if I typed it in) you could interface
> to the NR C/C++ routines.
>
> On Sun, 15 Jan 2006, Tolga Uzuner wrote:
>
> > Ah, my apologies. I meant the method outlined in section 10.5 of
> > Numerical Recipes using Powell's heuristic of discarding the direction
> > of largest decrease. A nice twist would be if linmin (line
> > minimization)  were to be implemented using the methods described in
> > 10.1-10.3, but that is just a nice-to-have.
> >
> > From what I understand, Nelder-Mead is the method in section 10.4
> > (downhill simplex), CG is section 10.6, and BFGS is section 10.7.
> >
> > Admittedly, optim is fantastic most of the time. However, I have a
> > factor model over hundreds of elements, where I am trying to back out
> > the factor distribution and it appears that the method in 10.5 might
> work.
> >
> > Thanks in advance in any case,
> > Tolga
> >
> > Prof Brian Ripley wrote:
> >
> >> (Michael) Powell (of Harwell and Cambridge) invented several
> >> optimization methods.  You will need to give us a precise reference.
> >>
> >> Powell is associated with earlier versions of the Nelder-Mead, CG and
> >> BFGS methods implemented in optim() but in each case different
> >> variants have best stood the course of time.
> >>
> >> On Sun, 15 Jan 2006, Tolga Uzuner wrote:
> >>
> >>> Has anyone implemented Powell's Method for minimisation in R ?
> >>
> >>
> >
> > ______________________________________________
> > [hidden email] mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide! http://www.R-project.org/posting-
> guide.html
> >
>
> --
> Brian D. Ripley,                  [hidden email]
> Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
> University of Oxford,             Tel:  +44 1865 272861 (self)
> 1 South Parks Road,                     +44 1865 272866 (PA)
> Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-
> guide.html

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Reply | Threaded
Open this post in threaded view
|

Re: Powell's Metod

Tolga Uzuner
Many thanks, let me look,
Tolga

Quoting Ravi Varadhan <[hidden email]>:

> Tolga,
>
> Have you considered an EM algorithm for your factor analysis problem?  A
> reference for this is:
>
> Rubin, D. and Thayer, D. (1982). EM algorithms for ML factor analysis.
> Psychometrika, 47(1):69--76.
>
> Hope this is helpful,
> Ravi.
>
> > -----Original Message-----
> > From: [hidden email] [mailto:r-help-
> > [hidden email]] On Behalf Of Prof Brian Ripley
> > Sent: Monday, January 16, 2006 9:31 AM
> > To: Tolga Uzuner
> > Cc: [hidden email]
> > Subject: Re: [R] Powell's Metod
> >
> > I am not aware of an R implementation of that method.  However, if you
> > meet the licensing conditions (which for my copy of NR do not allow me to
> > use the code on a Unix machine even if I typed it in) you could interface
> > to the NR C/C++ routines.
> >
> > On Sun, 15 Jan 2006, Tolga Uzuner wrote:
> >
> > > Ah, my apologies. I meant the method outlined in section 10.5 of
> > > Numerical Recipes using Powell's heuristic of discarding the direction
> > > of largest decrease. A nice twist would be if linmin (line
> > > minimization)  were to be implemented using the methods described in
> > > 10.1-10.3, but that is just a nice-to-have.
> > >
> > > From what I understand, Nelder-Mead is the method in section 10.4
> > > (downhill simplex), CG is section 10.6, and BFGS is section 10.7.
> > >
> > > Admittedly, optim is fantastic most of the time. However, I have a
> > > factor model over hundreds of elements, where I am trying to back out
> > > the factor distribution and it appears that the method in 10.5 might
> > work.
> > >
> > > Thanks in advance in any case,
> > > Tolga
> > >
> > > Prof Brian Ripley wrote:
> > >
> > >> (Michael) Powell (of Harwell and Cambridge) invented several
> > >> optimization methods.  You will need to give us a precise reference.
> > >>
> > >> Powell is associated with earlier versions of the Nelder-Mead, CG and
> > >> BFGS methods implemented in optim() but in each case different
> > >> variants have best stood the course of time.
> > >>
> > >> On Sun, 15 Jan 2006, Tolga Uzuner wrote:
> > >>
> > >>> Has anyone implemented Powell's Method for minimisation in R ?
> > >>
> > >>
> > >
> > > ______________________________________________
> > > [hidden email] mailing list
> > > https://stat.ethz.ch/mailman/listinfo/r-help
> > > PLEASE do read the posting guide! http://www.R-project.org/posting-
> > guide.html
> > >
> >
> > --
> > Brian D. Ripley,                  [hidden email]
> > Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
> > University of Oxford,             Tel:  +44 1865 272861 (self)
> > 1 South Parks Road,                     +44 1865 272866 (PA)
> > Oxford OX1 3TG, UK                Fax:  +44 1865 272595
> >
> > ______________________________________________
> > [hidden email] mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide! http://www.R-project.org/posting-
> > guide.html
>
>

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html