[SciPy-User] scipy.odr question

[Mogliii]

> I think you cannot tell it that the errors are exactly zero. However, to
> get least-squares result, you need to say that y-errors are way bigger
> than x-errors, for example
>
> 	sx=1e-99, sy=1.0
>
You are right, see attached image.

Now I will need to find out why peak-o-mat finds a better fit than mpfit
and leastsq... Also good to know that the uncertainty can be given as a
scalar and doesn't have to be a list.

Mogliii
-------------- next part --------------
A non-text attachment was scrubbed...
Name: odrfit.png
Type: image/png
Size: 52744 bytes
Desc: not available
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20120704/383970f5/attachment.png>