why does making all weights uniform produce a more desirable fit to the wing of the galaxy, at least to the eye?

0
Posted

why does making all weights uniform produce a more desirable fit to the wing of the galaxy, at least to the eye?

0

The main reason for this behavior is exactly that when more weight is given to bright pixels the central region is better fit than outer regions. This causes the sky to be fitted at a higher value, because the sky has one degree of freedom that can be used to minimize the residuals in the inner region. This causes the wings to be over-subtracted more than Poisson weighting, truncating the galaxy profile; in the extreme that the outer points don’t matter at all, the effect is the same thing as reducing the image fitting region. The other part of the reason is psychological: that most often when the data points (square points) lie above the model fit in the wings (solid line at R > 18 arcsec in this example), people will judge this to be evidence that the central region is over-weighted, because, as they reason, the model does not “see” the flux in the wing. Whereas, if the data points were to lie *below* the model, our inclination is to explain away the disagreement as there being a “

Related Questions