[SciPy-User] optimize.fmin_cg terminates when w - grad*1e-10 yields lower obj & grad
Mon Oct 5 13:51:27 CDT 2009
On Mon, Oct 5, 2009 at 2:40 PM, Jason Rennie <firstname.lastname@example.org> wrote:
> On Mon, Oct 5, 2009 at 2:35 PM, <email@example.com> wrote:
>> In this case, I either agree with Sebastian, or you are already at a
>> up to the usual precision, or your problem is badly scaled (my next
>> In your initial example, you had an improvement with a stepsize 1e-10,
>> however fmin_cg has a minimum stepsize of amin = 1e-8 (hardcoded) in
>> if my very fast skimming of the code is correct.
> Ah. This sounds like it may be the problem. When I use 1e-8, I don't get a
> smaller objective.
Then dividing your X or w by something large (1e5, 1e8) in the
should help ?
>> I don't know about fmin_cg but, if I remember correctly, I got this return
>> with other minimizers when I had an (almost) perfect fit, with no noise
>> in the
>> simulation. Otherwise, you could try to rescale your problem to make the
>> parameters larger in absolute value.
> Jason Rennie
> Research Scientist, ITA Software
> SciPy-User mailing list
More information about the SciPy-User