Cross-identification of Very Large Catalogues

Abstract

Modern astronomy has entered the era of very large catalogues, gathering information for over 10⁸ sources. Dedicated methods have been developed at the CDS to handle the huge amounts of data involved: powerful lossless compression, keeping direct access to the data on the basis of celestial position, allow fast queries on those very large catalogues.

We present the use of these very large datasets in the context of a data mining project undertaken by CDS and ESO. The challenge of cross-matching very large catalogues with other data sets is discussed. The question of likelihood of cross-identifications, using a statistical approach for large samples, is also addressed.