[daip] [!7112]: AIPS - AIPS JMFIT: Chi-sq

Eric Greisen do-not-reply at nrao.edu
Thu Aug 20 13:37:59 EDT 2015


Eric Greisen updated #7112
--------------------------

       Staff (Owner): Eric Greisen (was: -- Unassigned --)
              Status: Closed (was: Open)
                 Due: - Cleared - (was: 21 August 2015 03:16 PM)

AIPS JMFIT: Chi-sq
------------------

           Ticket ID: 7112
                 URL: https://help.nrao.edu/staff/index.php?/Tickets/Ticket/View/7112



It has been found that the mathematical methods of
estimating uncertainties in Gaussian fits of images in
radio astronomy do not tell the truth at all.  The
error bars formally out of JMFIT/IMFIT are usually 1 or
more orders of magnitude too small.  The error bars
quoted by JMFIT/IMFIT/SAD are based on studies done
by Condon et al. for the NVSS are are generally reasonable.
The estimate of whether the object is a point source is
also made qualitatively.  Read the explain file for details.

There should be some residual left even in a 100 sigma source.

Eric Greisen




------------------------------------------------------
Staff CP:  https://help.nrao.edu/staff



More information about the Daip mailing list