Good prior is very important

Another lesson,

Using vague prior may lead to completely wrong result. For example, suppose that we have a small galaxy with low signal-to-ratio. If we set priors of half-light radius of bulge and disk having large dynamic range (vague prior), bulge model may have unreasonably large half-light radius (larger than disk half-light radius) but high Sersic index to match the central part of galaxy. Then disk model also conspires to match the data in a wrong way. The “goodness of fit” might be as good as that of  “correct underlying model”. I found several such cases in 2MASS galaxy analysis.

For bright and large galaxies, this effect due to vague prior is less important. However as we go down to faint/small galaxies, we have to be more careful about our prior. Therefore I changed prior of 2MASS galaxy half-light radius to be within reasonable range: upper boundary is smaller than r_k20fe, a radius where surface brightness (mag/arcsec^2) is 20, which is about 1 sigma in the sky.


~ by ilsangyoon on February 7, 2009.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: