Re: [R] optim evils

2013-10-02 Thread Duncan Murdoch
On 02/10/2013 9:18 AM, Michael Meyer wrote: Greetings, In obedient deference to the demands of the collective I emailed a BUG report containing code and data to r-b...@r-project.org but found subsequently that I am unable to load the page http://bugs.r-project.org/ to check on the status of th

[R] optim evils

2013-10-02 Thread Michael Meyer
Greetings, In obedient deference to the demands of the collective I emailed a BUG report containing code and data to r-b...@r-project.org but found subsequently that I am unable to load the page http://bugs.r-project.org/ to check on the status of this report. Can anyone else load this page?

Re: [R] optim evils

2013-09-05 Thread Bert Gunter
Michael: Your parameter specification is probably over-determined, so that you have an infinite set of parameters that give essentially the same solution within numerical error. I would venture to guess that this will not be fixable with alternative optimizers. It is up to you to provide a sensibl

Re: [R] optim evils

2013-09-05 Thread Bert Gunter
Slight correction: On Thu, Sep 5, 2013 at 7:48 AM, Bert Gunter wrote: > Michael: > > Your parameter specification is probably over-determined, so that you have > an infinite set of parameter **values** that give essentially the same > solution within numerical error. I would venture to guess

Re: [R] optim evils

2013-09-05 Thread Martin Maechler
> Thanks for all replies. > The problem occurred in the following context: > A Gaussian one dimensional mixture (number of constituents, locations, > variances all unknown) > is to be fitted to data (as starting value to or in lieu of mixtools). A > likelihood maximization is performed. Cool.

[R] optim evils

2013-09-05 Thread Michael Meyer
Thanks for all replies. The problem occurred in the following context: A Gaussian one dimensional mixture (number of constituents, locations, variances all unknown) is to be fitted to data (as starting value to or in lieu of mixtools). A likelihood maximization is performed. I'll try to destill

[R] optim evils

2013-09-04 Thread Prof J C Nash (U30A)
in optimx package through other packages (bobyqa, nmkb, Rvmmin, Rcgmin) JN On 13-09-04 06:00 AM, r-help-requ...@r-project.org wrote: > Message: 67 > Date: Wed, 4 Sep 2013 16:34:54 +0800 (SGT) > From: Michael Meyer > To:"r-help@r-project.org" > Sub

Re: [R] optim evils

2013-09-04 Thread William Dunlap
Message- > From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On > Behalf > Of Michael Meyer > Sent: Wednesday, September 04, 2013 1:35 AM > To: r-help@r-project.org > Subject: [R] optim evils > > It would take some effort to extract selfconta

Re: [R] optim evils

2013-09-04 Thread Joshua Wiley
Hi Michael, You do not need to create a self-contained example from the mass of code where it is embedded, but given that optim() works in many cases, to file a bug report, you do need to give _an_ example where it is failing. Here is an example where it works great: > optim(1, fn = function(x)

[R] optim evils

2013-09-04 Thread Michael Meyer
It would take some effort to extract selfcontained code from the mass of code wherein this optimization is embedded. Moreover I would have to obtain permission from my employer to do so. This is not efficient. However some things are evident from the trace log which I have submitted: (a) L-BFGS-

Re: [R] optim evils

2013-09-03 Thread Rolf Turner
I don't think anyone can do much to help you unless you show us (a) your objective function "OF" and your starting value for "pars" --- which I do not see in your posting. Examples should be ***reproducible***!!! My personal experience with optim() has always been very good. cheers,

[R] optim evils

2013-09-03 Thread Michael Meyer
Greetings, I am in great anguish as the routine stats::optim shows unexplicable behaviour of various sorts. For one it is immune to the choice of optimization method and seems to always do the same. The following trace log N = 21, M = 5 machine precision = 2.22045e-16 At X0, 0 variables are ex