J/A+A/691/A223 Hot subdwarf binaries (Viscasillas Vazquez+, 2024)
Advanced classification of hot subdwarf binaries using artificial intelligence
techniques and Gaia DR3 data.
Viscasillas Vazquez C., Solano E., Ulla A., Ambrosch M., Alvarez M.A.,
Manteiga M., Magrini L., Santovena-Gomez R., Dafonte C., Perez-Fernandez E.,
Aller A., Drazdauskas A., Mikolaitis S., Rodrigo C.
<Astron. Astrophys. 691, A223 (2024)>
=2024A&A...691A.223V 2024A&A...691A.223V (SIMBAD/NED BibCode)
ADC_Keywords: Stars, subdwarf ; Optical
Keywords: methods: data analysis - techniques: spectroscopic -
binaries: general - subdwarfs
Abstract:
Hot subdwarf stars are compact blue evolved objects, that are located
by the blue end of the Horizontal Branch. Most models agree on a
common envelope binary evolution scenario in the Red Giant phase.
However, the current binarity rate for these objects is yet
unsolved.This study aims to develop a novel classification method for
identifying hot subdwarf binaries using Artificial Intelligence
techniques and data from the third Gaia data release (GDR3). The
methods used for hot subdwarf binary classification include supervised
and unsupervised machine learning techniques. Specifically, we have
used Support Vector Machines (SVM) to classify 3084 hot subdwarf stars
based on their colour-magnitude properties. Among these, 2815 objects
have Gaia DR3 BP/RP spectra, which were classified using
Self-Organizing Maps (SOM) and Convolutional Neural Networks (CNN).
Additional analysis onto a golden sample of 88 well-defined objects,
is also presented. The findings demonstrate a high agreement level
(∼70-90%) with the classifications from the Virtual Observatory Sed
Analyzer (VOSA) tool. SVM in a radial basis function achieves 70.97%
reproducibility for binary targets using photometry, and CNN reaches
84.94% for binary detection using spectroscopy. We also find that the
single-binary differences are especially observable on the infrared
flux in our Gaia DR3 BP/BR spectra, at wavelengths larger than ∼700nm.
We find that all the methods used are in fairly good agreement and are
particularly effective to discern between single and binary systems.
The agreement is also consistent with the results previously obtained
with VOSA. In global terms, considering all quality metrics, CNN is
the method that provides the best accuracy. The methods also appear
effective for detecting peculiarities in the spectra. While promising,
challenges in dealing with uncertain compositions highlight the need
for caution, suggesting further research is needed to refine
techniques and enhance automated classification reliability,
particularly for large-scale surveys.
Description:
Classification labels for the different methods used in this work,
where 0 means single and 1 means binary.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
tablea1.dat 78 2815 Classification labels for the different methods
--------------------------------------------------------------------------------
See also:
I/355 : Gaia DR3 Part 1. Main source (Gaia Collaboration, 2022)
Byte-by-byte Description of file: tablea1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 8 A8 --- --- [Gaia DR3]
10- 28 I19 --- GaiaDR3 Gaia DR3 identifier
30- 54 A25 --- Object Object identifier
56 I1 --- VOSA [0/1] Label from VOSA
58 I1 --- w-SVM-linear [0/1] Label from weighted SVM with
linear basis function
60 I1 --- w-SVM-rbf [0/1] Label from weighted SVM with
radial basis function
62 I1 --- SOM [0/1] Label from SOM
64 I1 --- CNN [0/1] Label from CNN
66- 70 F5.3 --- Pbinary-SOM [0/1] Probability that binary from SOM
72- 76 F5.3 --- Pbinary-CNN [0/1] Probability that binary from CNN
78 I1 --- Flag Number of agreeing labels from different
methods
--------------------------------------------------------------------------------
Acknowledgements:
Markus Ambrosch, markus.ambrosch(at)ff.vu.lt
(End) Patricia Vannier [CDS] 12-Oct-2024