comp.lang.idl-pvwave archive: archive » Re: weighted total least-squares regression

Home » Public Forums » archive » Re: weighted total least-squares regression

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: weighted total least-squares regression [message #68529]

Wed, 04 November 2009 13:51

natha
Messages: 482
Registered: October 2007

Senior Member

Thank you Wayne. I'm gonna try this last routine.
Regards,

nata

Report message to a moderator

Re: weighted total least-squares regression [message #68530 is a reply to message #68529]

Wed, 04 November 2009 12:54

wlandsman
Messages: 743
Registered: June 2000

Senior Member

On Nov 4, 2:14 pm, nata <bernat.puigdomen...@gmail.com> wrote:
> Hi gurus,
> I'm trying to find a routine to compute the weighted total least-
> squares regression.
> Does somebody know where can I found a routine to compute it ?
>
My understanding (from Wikipedia) is that weighted total least-squares
regression refers to fitting with errors in both coordinates.
(sixlin.pro is for fitting a straight line where the errors are not
given.) If you are just fitting a straight line you could try
fitexy ( http://idlastro.gsfc.nasa.gov/ftp/pro/math/fitexy.pro ) or
Craig Markwardt's linefitex ( http://www.physics.wisc.edu/~craigm/idl/down/linfitex.pro
) which uses MPFIT. These routines are different implementations of
the "Numerical Recipes" chapter.

But if you really want to do it right (though in a less
straightforward and much more computationally expensive way), then I
would use Brandon Kelley's Bayesian (Markov Chain Monte Carlo) program
linmix_err

http://idlastro.gsfc.nasa.gov/ftp/pro/math/linmix_err.pro

--Wayne

Report message to a moderator

Previous Topic:	weighted total least-squares regression
Next Topic:	Re: ENVI + IDL

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Oct 08 15:21:54 PDT 2025

Total time taken to generate the page: 0.00493 seconds