Séminaire/Seminar Galaxies |
| « Probabilistic source classification in the era of large photometric surveys » |
|
Meriam Ezziati |
|
The discovery of quasars at z > 7 in wide-field photometric surveys requires statistical methods capable of separating rare high-redshift sources from numerous contaminants such as brown dwarfs and intermediate-redshift galaxies. I will present Owl-z, a Bayesian classification tool developed for the probabilistic selection of high-redshift quasars in the Euclid Wide Survey.
Owl-z computes posterior probabilities by combining multi-band photometry with prior models of quasars and contaminants, enabling a controlled trade-off between completeness and purity through an adjustable probability threshold. Using Euclid simulations spanning 7 = z = 12 and 18 = Hmag = 24.5, Owl-z achieves high completeness and purity for bright sources with limited redshift dependence. Its application to Euclid observations demonstrates robust performance on real data and effective characterization of contaminants. Owl-z provides a transparent and adaptable probabilistic framework for efficient high-redshift quasar selection in large photometric surveys. |
|
jeudi 19 février 2026 - 11:30 Salle Entresol Daniel Chalonge Institut d'Astrophysique de Paris |
| Page web du séminaire / Seminar's webpage |