[SciPy-User] optimize.fmin_cg terminates when w - grad*1e-10 yields lower obj & grad

josef.pktd@gmai... josef.pktd@gmai...
Mon Oct 5 13:51:27 CDT 2009


On Mon, Oct 5, 2009 at 2:40 PM, Jason Rennie <jrennie@gmail.com> wrote:
> On Mon, Oct 5, 2009 at 2:35 PM, <josef.pktd@gmail.com> wrote:
>>
>> In this case, I either agree with Sebastian, or you are already at a
>> minimum
>> up to the usual precision, or your problem is badly scaled (my next
>> guesses).
>>
>> 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
>> linesearch
>> 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.
> Jason

Then dividing your X or w by something large (1e5, 1e8) in the
objective function
should help ?

Josef

>
>>
>> I don't know about fmin_cg but, if I remember correctly, I got this return
>> code
>>  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.
>>
>> Josef
>
> --
> Jason Rennie
> Research Scientist, ITA Software
> 617-714-2645
> http://www.itasoftware.com/
>
>
> _______________________________________________
> SciPy-User mailing list
> SciPy-User@scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
>
>


More information about the SciPy-User mailing list