Inspecting the spatial fit residuals

What you will learn

You will learn how to use the csresmap script to inspect the spatial fit residuals.

After having done a maximum likelihood fit it is good practice to inspect the fit residuals. You do this with the csresmap script that creates a residual sky map of the events from which the fitted model components were subtracted. You run the csresmap script as follows:

$ csresmap
Input event list, counts cube, or observation definition XML file [events.fits] cntcube.fits
Input model cube file (generated with ctmodel) [NONE]
Input exposure cube file [NONE] expcube.fits
Input PSF cube file [NONE] psfcube.fits
Input background cube file [NONE] bkgcube.fits
Input model definition XML file [$CTOOLS/share/models/crab.xml] crab_results.xml
Residual map computation algorithm (SUB|SUBDIV|SUBDIVSQRT|SIGNIFICANCE) [SIGNIFICANCE]
Output residual map file [resmap.fits]

Note

csresmap is a Python script while the other tools that you have used so far are C++ binary executables. From the User perspective, Python scripts and C++ binary executables behave the same way, but to distinguish both all Python scripts names start with cs while all C++ binary executables names start with ct.

The csresmap script produces the FITS file resmap.fits that contains the residual counts map. The image is displayed below using ds9 with a linear color scale and with a 3 pixel Gaussian kernel smoothing applied. Obviously, there are no significant residuals, which indicates that the model fit was satisfactory.

../../../../_images/residual_map.png

Residual sky map of the selected events with all model components subtracted

Note

csresmap implements different algorithms for the computation of the residuals. These are:

  • SUB: \(DATA - MODEL\)

  • SUBDIV: \((DATA - MODEL) / MODEL\)

  • SUBDIVSQRT: \((DATA - MODEL) / \sqrt{MODEL}\)

  • SIGNIFICANCE: \({\rm sign}(DATA-MODEL) \times \sqrt{ 2 \times ( DATA \times \ln \left(\frac{DATA}{MODEL} \right) + MODEL - DATA ) }\)

By default the SIGNIFICANCE algorithm is used.