[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Gaussian fit of a peak
From: |
Rob Mahurin |
Subject: |
Re: Gaussian fit of a peak |
Date: |
Wed, 24 Sep 2008 11:26:23 -0400 |
On Sep 24, 2008, at 4:04 AM, Andrea Cimatoribus wrote:
Unfortunately, I am not sure I will be able to keep "central" area
only, since I am fitting something that resembles a phonon spectra.
In fact, I have multiple peakes to the left and right with respect
to the "zero", but fitting everything with a gaussian balanced to
the zero energy pretty well. I need to do this in order to sum
spectra together. I will let you know if and how I solve this
problem if it is of any intereset.
AC
The leasqr wrapper I mentioned can handle multiple overlapping
Gaussians. You might wonder, for instance, whether a fit to two
unresolved peaks gives a smaller chi-squared per degree of freedom
than a fit to a single peak. This is what the "demo" does for a
dodgy problem with two exponentials. (Though you have to find the
chi2 yourself, sum( ((y-f5)./errscale).^2 ), after the demo runs,
because I wasn't thinking of this when I wrote it).
If you have some constant background you'll have to subtract it by
hand for now.
code, again, at
http://sns.phys.utk.edu/~mahurin/octave/
Cheers,
Rob
--
Rob Mahurin
Dept. of Physics & Astronomy
University of Tennessee phone: 865 207 2594
Knoxville, TN 37996 email: address@hidden
Re: Gaussian fit of a peak, Rob Mahurin, 2008/09/23
Re: Gaussian fit of a peak, Julius Smith, 2008/09/23