[Ifeffit] Spline range?

[Matt Newville]

Wei,

On Sun, Sep 18, 2011 at 9:10 AM, Wei Li <weili at udel.edu> wrote:
> Dear all,
>
> I am wondering about what range should be chosen in Athena? Some
literature
> says from 1 to the end; some says from 0.5 to the end.

What literature says this? Do you mean to imply that someone is saying it
should be 0.5 or 1.0, and *not* some other value?

In principle, the background should start at k=0, and one should use
k-weight=1, as
Integral_E=E0^E=infinity chi(E) dE = 0
Integral_k=0^k=infinity chi(k) * k * dk = 0

In practice, that is nearly useless beyond the suggestion that the integral
(which is to say, the DC component), *should* be zero.

Using kmin between 0 and 1 and k-weight between 0 and 2 usually "work best",
especially in the sense of "stable". In my experience, the value of E0 is
much more important than kmin, and it is hard to predict whether k-weight =
1 or 2 works better.

My advice is to not over-think or spend too much time worrying about
background subtraction.

--Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20110919/1f68b675/attachment.html>