Bayesian approach for matching multiple stellar observations

The cross-identification of sources in separate catalogs is one of the most basic tasks in observational astronomy. Recently Budavari and Szalay (2008) formulated the problem in the probability theory, and laid down the statistical foundations of an extendable methodology. An application of the same Bayesian approach to stars is presented that, we know, can measurably move on the sky, and it is shown how to associate their observations. Models are studied on a sample of stars in the Sloan Digital Sky Survey, which allow for an unknown proper motion per object, and their improvements are shown over the simpler static model. The new models and conclusions are directly applicable to the upcoming surveys, whose data sets will be most likely dominated by stars in the region of the Galactic Plane.