| Re: fitting with asymmetric errors? [message #74279] |
Fri, 07 January 2011 10:37 |
pgrigis
Messages: 436
Registered: September 2007
|
Senior Member |
|
|
Hi,
if the errors are asymmetric, it means that the
underlying distribution is non-gaussian.
Now most of the machinery used for error propagation,
fitting, etc. assumes gaussian errors.
Therefore, I would think you may need to move beyond
those simple methods if you want to get good results.
You may have to use an approach that models both data
and errors.
Ciao,
Paolo
On Jan 7, 11:46 am, Elizabeth <bartlett....@gmail.com> wrote:
> Hi all,
> I'm trying to fit a curve to a data set with asymmetric errors. I
> usually use mpfitexpr when fitting, but cannot see a way to input
> anything other than symmetric errors (though this is could well be a
> problem with my understanding rather than the routine!). If mpfitexpr
> is not the task for the job, is there another routine you would
> suggest?
>
> Thank you very much for your time! :)
> Liz
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
