[SciPy-User] ODR fitting several equations to the same parameters
Fri Nov 13 11:21:14 CST 2009
On Fri, Nov 13, 2009 at 12:18 PM, <email@example.com> wrote:
> On Fri, Nov 13, 2009 at 11:44 AM, ms <firstname.lastname@example.org> wrote:
>> email@example.com ha scritto:
>>> On Fri, Nov 13, 2009 at 10:28 AM, ms <firstname.lastname@example.org> wrote:
>>>> email@example.com ha scritto:
>>>>> On Thu, Nov 12, 2009 at 10:04 AM, ms <firstname.lastname@example.org> wrote:
>>>>>> email@example.com ha scritto:
>>>>> an example
>>>>> (quickly written and not optimized, there are parts I don't remember
>>>>> about curve_fit, fixed parameters could be better handled by a class)
>>>> Hmm, it seems I don't have curve_fit -I am constrained to use
>>>> scipy-0.6.0 and there's no chance to change that (it's an external server).
>>> You can just copy the function (plus 2 helper functions) from the
>>> current trunk. You would need to add the imports. Alternatively you
>>> can just use optimize.leastsq directly, using curve_fit as a recipe.
>> A further question: It seems to me it works only if the data sets have
>> the same size, because what gets minimized is then the matrix. What
>> about datasets with different sizes?
> In the example, I just did the stacking based on the 2d array to have
> it quickly written, for unequal sized data groups it is easier to work
> directly with the stacked array, and just index into it, or for
> example create a `b` array that has the values repeated corresponding
> to the group sizes. (Same story as with balanced versus unbalance
> Do you have a non-linear ODR example? I didn't even know ODR can do
> non-linear parameter estimation.
Or maybe I knew it and just forgot about it.
>> SciPy-User mailing list
More information about the SciPy-User