Re: Curve fit: strange documentation (josef.pktd@gmail.com)

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: Curve fit: strange documentation (josef.pktd@gmail.com)

federico vaggi-2
In weighted least squares, the residuals are weighted by their standard deviation - so it's basically the same thing.  



_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: Curve fit: strange documentation (josef.pktd@gmail.com)

josef.pktd


On Fri, Feb 7, 2014 at 1:37 PM, federico vaggi <[hidden email]> wrote:
In weighted least squares, the residuals are weighted by their standard deviation - so it's basically the same thing.  


 
It'a not a question whether they are "basically the same thing", it's a question of what the (imprecise) terms in the docstring actually mean.
 
`WLS-weights` = 1/sigma**2
but the transformation uses
w = 1/ sigma
or maybe weights just refers to general weighting, not to WLS-weights (natural language usage instead of formal definition)
 
Is it squared or not, and is it the inverse or not?
 
sigma is the standard deviation, except it's only proportional because we scale it (by default) with the residual standard deviation.
 
Josef
INALL
 

_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user



_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user