Automated stellar classification for large surveys: a review of methods and results

Abstract

Current and future large astronomical surveys will yield multiparameter databases on millions or even billions of objects. The scientific exploitation of these will require powerful, robust, and automated classification tools tailored to the specific survey. Partly motivated by this, the past five to ten years has seen a significant increase in the amount of work focused on automated classification and its application to astronomical data. In this article, I review this work and assess the current status of automated stellar classification, with particular regard to its potential application to large astronomical surveys. I examine both the strengths and weaknesses of the various techniques and how they have been applied to different classification and parametrization problems. I finish with a brief look at the developments still required in order to apply a stellar classifier to a large survey.