Multivariate Approaches to Classification in Extragalactic Astronomy

Didier Fraix-Burnet (1) , Marc Thuillard (2) , Asis Kumar Chattopadhyay (3)

1 IPAG - Institut de Planétologie et d'Astrophysique de Grenoble
2 La Colline
3 University of Calcutta

Clustering objects into synthetic groups is a natural activity of any science. Astrophysics is not an exception and is now facing a deluge of data. For galaxies, the one-century old Hubble classification and the Hubble tuning fork are still largely in use, together with numerous mono-or bivariate classifications most often made by eye. However, a classification must be driven by the data, and sophisticated multivariate statistical tools are used more and more often. In this paper we review these different approaches in order to situate them in the general context of unsupervised and supervised learning. We insist on the astrophysical outcomes of these studies to show that multivariate analyses provide an obvious path toward a renewal of our classification of galaxies and are invaluable tools to investigate the physics and evolution of galaxies.

Mots clés: Multivariate analysis.; Galaxies; phylogenetic methods; clustering; Classification

Domaines: Cosmologie et astrophysique extra-galactique [astro-ph.CO]; Applications [stat.AP]; Bio-Informatique, Biologie Systémique [q-bio.QM]; Astrophysique [astro-ph]

Didier Fraix-Burnet, Marc Thuillard, Asis Kumar Chattopadhyay. Multivariate Approaches to Classification in Extragalactic Astronomy. Frontiers in Astronomy and Space Sciences, 2015, 2 (3), pp.00. ⟨10.3389/fspas.2015.00003⟩. ⟨hal-01184937⟩