J/ApJS/228/24 GALAH semi-automated classification scheme (Traven+, 2017)
The GALAH survey: classification and diagnostics with t-SNE reduction of
spectral information.
Traven G., Matijevic G., Zwitter T., Zerjal M., Kos J., Asplund M.,
Bland-Hawthorn J., Casey A.R., De Silva G., Freeman K., Lin J.,
Martell S.L., Schlesinger K.J., Sharma S., Simpson J.D., Zucker D.B.,
Anguiano B., Da Costa G., Duong L., Horner J., Hyde E.A., Kafle P.R.,
Munari U., Nataf D., Navin C.A., Reid W., Ting Y.-S.
<Astrophys. J. Suppl. Ser., 228, 24-24 (2017)>
=2017ApJS..228...24T 2017ApJS..228...24T (SIMBAD/NED BibCode)
ADC_Keywords: Surveys ; Milky Way ; Stars, bright
Keywords: binaries: general; catalogs; methods: data analysis; stars: activity;
stars: peculiar; surveys
Abstract:
Galah is an ongoing high-resolution spectroscopic survey with the goal
of disentangling the formation history of the Milky Way using the
fossil remnants of disrupted star formation sites that are now
dispersed around the Galaxy. It is targeting a randomly selected
magnitude-limited (V≤14) sample of stars, with the goal of observing
one million objects. To date, 300000 spectra have been obtained. Not
all of them are correctly processed by parameter estimation pipelines,
and we need to know about them. We present a semi-automated
classification scheme that identifies different types of peculiar
spectral morphologies in an effort to discover and flag potentially
problematic spectra and thus help to preserve the integrity of the
survey results. To this end, we employ the recently developed
dimensionality reduction technique t-SNE (t-distributed stochastic
neighbor embedding), which enables us to represent the complex
spectral morphology in a two-dimensional projection map while still
preserving the properties of the local neighborhoods of spectra. We
find that the majority (178483) of the 209533 Galah spectra considered
in this study represents normal single stars, whereas 31050 peculiar
and problematic spectra with very diverse spectral features pertaining
to 28579 stars are distributed into 10 classification categories: hot
stars, cool metal-poor giants, molecular absorption bands, binary
stars, Hα/Hβ emission, Hα/Hβ emission
superimposed on absorption, Hα/Hβ P-Cygni, Hα/Hβ
inverted P-Cygni, lithium absorption, and problematic. Classified
spectra with supplementary information are presented in the catalog,
indicating candidates for follow-up observations and population
studies of the short-lived phases of stellar evolution.
Description:
The GALactic Archaeology with HERMES (GALAH) survey was the main
driver for the construction of Hermes (High Efficiency and Resolution
Multi-Element Spectrograph), a fiber-fed multi-object spectrograph on
the 3.9m Anglo-Australian Telescope. Its spectral resolving power (R)
is about 28000, and there is also an R=45000 mode using a slit mask.
Hermes has four simultaneous non-contiguous spectral arms centered at
4800, 5761, 6610, and 7740Å, covering about 1000Å in total,
including Hα and Hβ lines. About 300000 spectra have been
taken to date, including various calibration exposures. However, we
concentrate on ∼210000 spectra recorded before 2016 January 30.
We devise a custom classification procedure which is based on two
independently developed methods, the novel dimensionality reduction
technique t-SNE (t-distributed stochastic neighbor embedding; van der
Maaten & Hinton 2008, Journal of Machine Learning Research 9, 2579)
and the renowned clustering algorithm DBSCAN (Ester+ 1996, Proc. 2nd
Int. Conf. on KDD, 226 ed. E. Simoudis, J. Han, and U. Fayyad).
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table1.dat 163 73 *Classification categories based on the general
projection map
table3.dat 163 39 *Classification categories based on the specific
projection map produced in the search for
young/active stars
table4.dat 435 12210 Catalog containing results of our classification
refs.dat 217 464 References listed in column "ADS" of table 4;
column converted in table by CDS
--------------------------------------------------------------------------------
Note on table1.dat: This table lists six distinct categories that were defined
using the classification procedure. This classification is not strictly
limited to peculiar objects that have spectra without a counterpart in the
library of synthetic spectra, although they remain the principal motivation
for this work. It is instead a search for any coherent group in the
projection map, from which a category of interest can be selected.
See section 4 for further explanations.
Note on table3.dat: We also present additional classification results based on
a more specific projection map, in contrast to the general map presented in
Section 4. These results follow the same procedure, but with different t-SNE
input parameters and input spectral ranges.
See section 5 for further explanations.
--------------------------------------------------------------------------------
See also:
B/wds : The Washington Visual Double Star Catalog (Mason+ 2001-2014)
B/sb9 : 9th Catalogue of Spectroscopic Binary Orbits (Pourbaix+ 2004-2014)
I/337 : Gaia DR1 (Gaia Collaboration, 2016)
V/146 : LAMOST DR1 catalogs (Luo+, 2015)
II/328 : AllWISE Data Release (Cutri+ 2013)
II/312 : GALEX-DR5 (GR5) sources from AIS and MIS (Bianchi+ 2011)
J/MNRAS/465/3203 : GALAH observational overview (Martell+, 2017)
J/A+A/581/A52 : Gaia-ESO Survey: Hα emission stars (Traven+, 2015)
J/ApJ/808/16 : The Cannon: new approach to determine abundances (Ness+, 2015)
J/AcA/63/21 : VI light curves of Galactic LPVs (Soszynski+, 2013)
J/AJ/140/184 : RAVE double-lined spectroscopic binaries (Matijevic+, 2010)
http://galah-survey.org/ : GALAH home page
Byte-by-byte Description of file: table[13].dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 48 A48 --- Cat Classification category (1)
50- 53 I4 --- o_Cat [18/4130] Number of sources in Cat
55-103 A49 --- MType SIMBAD main source type (2)
105-107 I3 --- o_MType [0/371] Number of sources in MType
109-157 A49 --- OType SIMBAD other source type (2)
159-163 I5 --- o_OType [0/1486] Number of sources in OType
--------------------------------------------------------------------------------
Note (1): Catalog excludes the results for spectra from the Problematic
category, since these mainly stand out for data reduction reasons
and will be recoverable in the upgraded versions of the reduction
pipeline.
Note (2): SIMBAD defines a main type for each astronomical object in its
database, and usually several other types generally inferred from
its identifiers. For these columns the less interesting type "Star"
is excluded.
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table4.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 5 I5 --- Seq [0/12209] Internal catalog index number (1)
7- 13 I7 --- GALAH [0/9520722] GALAH unique identifier
15- 27 F13.7 d MJD Modified Julian Date of observation
29- 44 F16.12 deg RAdeg Right Ascension in decimal degrees (J2000)
46- 63 F18.14 deg DEdeg Declination in decimal degrees (J2000)
65- 86 A22 --- GClass General classification category (see table 1)
88- 135 A48 --- SClass Specific classification category (see table 3)
137- 164 A28 --- Simbad SIMBAD identifier for 3289 sources
166- 173 F8.6 arcsec Sep [0/1]? Angular distance from Galah target
to SIMBAD source
175- 223 A49 --- MType Main SIMBAD type
225- 382 A158 --- OType Other SIMBAD type
384 I1 --- nRad [0/1] Number of VizieR tables for radio range
386- 387 I2 --- nIR [0/31] Number of VizieR tables for IR range
391- 392 I2 --- nOpt [0/61] Number of VizieR tables for optical range
396 I1 --- nUV [0/2] Number of VizieR tables for UV range
401 I1 --- nEUV [0] Number of VizieR tables for EUV range
406- 407 I2 --- nXRay [0/19] Number of VizieR tables for X-ray range
411 I1 --- nGam [0] Number of VizieR tables for Gamma-ray range
416- 435 A20 --- OClass OGLE variable star type/class
--------------------------------------------------------------------------------
Note (1): Catalog excludes the results for spectra from the Problematic
category, since these mainly stand out for data reduction reasons
and will be recoverable in the upgraded versions of the reduction
pipeline.
--------------------------------------------------------------------------------
Byte-by-byte Description of file: refs.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 5 I5 --- Seq [144/12188] Internal catalog index number
7- 25 A19 --- BibCode ADS bibcode
27- 217 A191 --- Title Title of the reference
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Prepared by [AAS], Emmanuelle Perret [CDS] 13-Apr-2017