Astrophysics with Radioactive Isotopes

Barium stars classification with s-process models using machine learning techniques

Not scheduled

20m

Oral Presentation

Speaker

Jacqueline den Hartogh

Description

On Monday, Andrés Yagüe López already presented our research on using machine learning techniques to compare models and observations of CEMP stars. In this talk, we will present our results on Ba stars (submitted to A&A). This analysis is done by systematically comparing a database of 169 Ba stars (de Castro et al 2016) with Asymptotic Giant Branch stellar models from FRUITY (Cristallo et al. 2016) and Monash (Karakas et al, 2018). The elements that we use in the classification are selected by systematically removing s-process elements from our AGB models to identify the elements whose removal has the biggest positive effect on the classification. Our final set includes: Fe, Rb, Sr, Zr, Ru, Nd, Ce, Sm, and Eu and excludes Nb, Y, Mo, and La. Our algorithms found final classifications for 166 of the 169 Ba stars in our sample. The classifications of both sets of AGB final surface abundances show similar distributions and both peak at 2.25 Mo and Z = 0.01. We also investigated why the removal of Nb, Y, Mo, and La improves our classification and identified 43 stars for which the exclusion had the biggest effect. These stars have a statistically significant different abundances for Mo, Nb, La and Sm than the other Ba stars in our sample. We discuss potential reasons for these differences in abundance patterns.