A comparative analysis of data mining algorithms to mitigate spurious detections in Gaia

Abstract

Gaia is an ESA mission that observes about 50 million sources per day. A small part of these detections are considered spurious generated for example by cosmic rays. The main objective of this study is to perform a comparative analysis of several algorithms to automatically detect spurious detections. Successfully identifying these detections is important to prevent them from entering the cross-match stage where they create several problems and degrade resolution performance. We will use appropriate metrics to determine the execution and assess the algorithms. Finally, it will be discussed if any of these data mining algorithms could be a good solution to the spurious detection problem.