How problematic are parameter degeneracies?
Lastly, dire fears about parameter degeneracies being a problem for large numbers of free parameters and crowding did not come true in these simulations. Recall that about 50% of all objects have neighbors which need to be fitted together or masked out. Moreover, in all situations where simultaneous fitting is necessary the initial parameters of all components involved would be quite crude. If parameter degeneracy were a serious issue these are the perfect circumstances to see the effects. In fact, as shown in Figures 3 and 4, GALFIT produces an overall scatter in the parameters (size, luminosity, Sérsic index) that compares very well with GIM2D single component fits. In single component fits, parameter degeneracy is virtually a non-issue, on top of the fact that GIM2D is a Metropolis-based algorithm (however, see Boris’ cautions on proper use of GIM2D).
Related Questions
- Ive been using the IsNull function but it seems to be ignoring the second parameter (just returning True or False instead). Is this function supported and/or is there an alternative?
- My PSM-4900 and PSM-2100 do not agree on the Eb/No parameter with Reed-Solomon codec installed?
- How problematic are parameter degeneracies?
