[SciPy-user] nonlinear fit with non uniform error?

massimo sandal massimo.sandal@unibo...
Thu Jun 21 07:31:24 CDT 2007


Matthieu Brucher ha scritto:
>     I guess you refer to the distribution of the error for *each single
>     point*, not the distribution of the average error in the dataset for
>     different points. In this case yes, it is Gaussian, so there should be
>     no problem.

> It is the distribution for all errors. If it is the same distribution 
> for all points, OK with least squares. If it is not, you have to scale 
> the points so that the errors follow the same gaussian law.

Sigh, there is some misunderstanding between us (surely due to my utter 
ignorance).

Imagine I have three data points, A B C.
Usually, if the error is uniform (on Y) can be:
Ay +/- 5
By +/- 5
Cy +/- 5

In my case it is:
Ay +/- 5
By +/- 7
Cy +/- 11

Now, the error in all those three cases behaves gaussianly, but with 
different widths.

1)Does this mean that least squares is NOT ok?
2)What does "rescaling" mean in this context?

m.

-- 
Massimo Sandal
University of Bologna
Department of Biochemistry "G.Moruzzi"

snail mail:
Via Irnerio 48, 40126 Bologna, Italy

email:
massimo.sandal@unibo.it

tel: +39-051-2094388
fax: +39-051-2094387
-------------- next part --------------
A non-text attachment was scrubbed...
Name: massimo.sandal.vcf
Type: text/x-vcard
Size: 274 bytes
Desc: not available
Url : http://projects.scipy.org/pipermail/scipy-user/attachments/20070621/25bcb81b/attachment.vcf 


More information about the SciPy-user mailing list