J/ApJS/273/34 YSOs from LAMOST DR9 using deep learning (Tan+, 2024)
A robust young stellar object identification method based on deep learning.
Tan L., Liu Z., Wang X., Mei Y., Wang F., Deng H., Liu C.
<Astrophys. J. Suppl. Ser., 273, 34 (2024)>
=2024ApJS..273...34T 2024ApJS..273...34T
ADC_Keywords: YSOs; Surveys; Spectra, optical
Keywords: Young stellar objects ; Stellar spectral types ;
Stellar classification ; Astronomy data analysis ; Neural networks
Abstract:
Young stellar objects (YSOs) represent the earliest stage in the
process of star formation, offering insights that contribute to the
development of models elucidating star formation and evolution. Recent
advancements in deep-learning techniques have enabled significant
strides in identifying special objects within vast data sets. In this
paper, we present a YSO identification method based on deep-learning
principles and spectra from the LAMOST. We designed a structure based
on a long short-term memory network and a convolutional neural network
and trained different models in two steps to identify YSO candidates.
Initially, we trained a model to detect stellar spectra featuring the
Hα emission line, achieving an accuracy of 98.67%. Leveraging
this model, we classified 10,495,781 stellar spectra from LAMOST,
yielding 76,867 candidates displaying a Hα emission line.
Subsequently, we developed a YSO identification model, which achieved
a recall rate of 95.81% for YSOs. Utilizing this model, we further
identified 35,021 YSO candidates from the Hα emission-line
candidates. Following cross validation, 3204 samples were identified
as previously reported YSO candidates. We eliminated samples with low
signal-to-noise ratios and M dwarfs by using the equivalent widths of
the NII and HeI emission lines and visual inspection, resulting in a
catalog of 20,530 YSO candidates. To facilitate future research
endeavors, we provide the obtained catalogs of Hα emission-line
star candidates and YSO candidates along with the code used for
training the model.
Description:
We utilized spectra sourced from the LAMOST DR9 catalog for data-set
construction.
Given the absence of young stellar object (YSO) classifications in
LAMOST's data, we draw upon YSOs previously documented by researchers
to compile our YSO catalog, which were derived from
Gutermuth+ (2009, J/ApJS/184/18), Megeath+ (2012, J/AJ/144/192),
Rapson+ (2014, J/ApJ/794/124), Marton+ (2016, J/MNRAS/458/3479 and
2019, II/360), Vioque+ (2020, J/A+A/638/A21),
Cornu & Montillaud (2021, J/A+A/647/A116), Kuhn+ (2021, J/ApJS/254/33),
McBride+ (2021, J/AJ/162/282), Sanchez-Saez+ (2021AJ....161..141S 2021AJ....161..141S),
Wilson+ (2023, VII/293), Zhang+ (2023, J/ApJS/267/7), and
Rimoldini+ (2023A&A...674A..14R 2023A&A...674A..14R). See Section 2.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table2.dat 115 76867 Basic information of the 76867 Hα emission
line star candidates
table4.dat 96 20530 Basic information of 20530 YSO candidates
--------------------------------------------------------------------------------
See also:
II/360 : Gaia DR2 x AllWISE catalogue (Marton+, 2019)
V/156 : LAMOST DR7 catalogs (Luo+, 2019)
VII/293 : Classifier catalogue of class II YSO posteriors (Wilson+, 2023)
J/ApJ/667/308 : Weak-line T Tauri in Spitzer c2d Survey. II. (Cieza+, 2007)
J/ApJ/688/1142 : Star formation in W5: Spitzer observations (Koenig+, 2008)
J/A+A/504/461 : YSOs in L1630N and L1641 (Fang+, 2009)
J/ApJS/184/18 : Spitzer survey of young stellar clusters (Gutermuth+, 2009)
J/ApJS/186/259 : Taurus Spitzer survey: new candidate members (Rebull+, 2010)
J/AJ/144/192 : Spitzer survey of Orion A and B. I. YSOs (Megeath+, 2012)
J/ApJ/794/124 : Star forming region NGC2264 Spitzer sources (Rapson+, 2014)
J/ApJ/827/96 : WISE census of YSOs in Canis Major (Fischer+, 2016)
J/ApJS/224/5 : Herschel Orion Protostar Survey (HOPS): SEDs (Furlan+, 2016)
J/MNRAS/458/3479 : SVM selection of WISE YSO Candidates (Marton+, 2016)
J/A+A/638/A21 : Herbig Ae/Be and classical Be stars catalog (Vioque+, 2020)
J/A+A/647/A116 : YSO candidate catalog from ANN (Cornu+, 2021)
J/ApJS/254/33 : Gal. midplane Spitzer/IRAC cand. YSOs (SPICY) (Kuhn+, 2021)
J/AJ/162/282 : Untangling Galaxy. III. PMS stars (Mcbride+, 2021)
J/ApJS/267/7 : YSO candidates from LAMOST LRS DR9 & ZTF (Zhang+, 2023)
Byte-by-byte Description of file: table2.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 19 A19 --- LAMOST LAMOST designation (JHHMMSS.ss+DDMMSS.s)
21- 30 F10.6 deg RAdeg LAMOST Right Ascension (J2000)
32- 40 F9.6 deg DEdeg [-9.7/88.6] LAMOST declination (J2000)
42- 50 I9 --- ObsID [101094/914312180] LAMOST observation
identifier
52- 59 F8.2 --- SNR-g [0/999]?=-9999 LAMOST Signal-to-Noise in the
g-band
61- 68 F8.2 --- SNR-u [0/499]?=-9999 LAMOST Signal-to-Noise in the
u-band
70- 77 F8.2 --- SNR-z [0/999]?=-9999 LAMOST Signal-to-Noise in the
z-band
79- 86 F8.2 --- SNR-r [0/999]?=-9999 LAMOST Signal-to-Noise in the
r-band
88- 95 F8.2 --- SNR-i [0/999]?=-9999 LAMOST Signal-to-Noise in the
i-band
97- 115 A19 --- SClass Simbad cross-validation classification
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table4.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 19 A19 --- LAMOST LAMOST designation (JHHMMSS.ss+DDMMSS.s)
20 A1 --- f_LAMOST [*] Li I absorption line were observed in
LAMOST
22- 31 F10.6 deg RAdeg LAMOST Right Ascension (J2000)
33- 41 F9.6 deg DEdeg [-9.2/86.4] LAMOST declination (J2000)
43- 51 I9 --- ObsID [304080/914213134] LAMOST observation
identifier
53- 60 F8.2 --- SNR-g [0/908]?=-9999 LAMOST Signal-to-Noise in the
g-band
62- 69 F8.2 --- SNR-u [0/261]?=-9999 LAMOST Signal-to-Noise in the
u-band
71- 78 F8.2 --- SNR-z [0/999]?=-9999 LAMOST Signal-to-Noise in the
z-band
80- 87 F8.2 --- SNR-r [0/585]?=-9999 LAMOST Signal-to-Noise in the
r-band
89- 96 F8.2 --- SNR-i [0/999]?=-9999 LAMOST Signal-to-Noise in the
i-band
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Prepared by [AAS], Emmanuelle Perret [CDS] 08-Jan-2025