[AstroPy] Maximum likelihood fit to unbinned data

[Eduardo Balbinot]

Hi there,

thanks for the references. I managed to solve the issue using the reference
Paul sent.

I was trying to fit a density profile to a distribution of star. The usual
way of doing it is to build a histogram of dN/dR (number of points/radial
bins). But that introduces biases due to the binning choice.

A way to avoid binning is to use a likelihood approach. So I followed this
paper 2008ApJ...684.1075M.

If anyone is interested in the code I wrote for that please ask!

Cheers and thanks again.
-- Eduardo Balbinot


On Sat, Dec 22, 2012 at 9:14 PM, Paul Barrett <pebarrett at gmail.com> wrote:

> Eduardo,
>
> What type of data are you trying to fit; time series, spectral or image?
>
> You might want to take a look at what the gamma-ray community has
> done, since almost all of their analysis involves photon lists.  You
> could also try searching the astrostatistics website.
>
> Personally, I would use a Bayesian approach.  I have successfully used
> the Nested Sampling approach to fit FUSE and Chandra spectra. This
> algorithm succeeded where the traditional Levenberg-Marquardt
> algorithm did not.  The algorithm is described in "Data Analysis: A
> Bayesian Approach" by D.S. Sivia.  One of the best introductions to
> the subject in my opinion.
>
>  -- Paul
>
> On Sat, Dec 22, 2012 at 1:22 PM, Eduardo Balbinot
> <eduardo.balbinot at gmail.com> wrote:
> > A bit off-topic, but....
> >
> > Anyone could point to a good guide to fitting unbinned data using a
> maximum
> > likelihood estimator?
> >
> > It would be a nice idea to have it on astropy with most used density
> > profiles (Hubble, Plummer, King, etc...). If I manage to do it we could
> > discuss including it on astropy.
> >
> > Thanks!
> > -- Eduardo Balbinot
> >
> > _______________________________________________
> > AstroPy mailing list
> > AstroPy at scipy.org
> > http://mail.scipy.org/mailman/listinfo/astropy
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/astropy/attachments/20121226/75f1b048/attachment.html>