J/A+A/475/1159 Supervised classification of variable stars (Debosscher+, 2007)
Automated supervised classification of variable stars. I. Methodology.
Debosscher J., Sarro L.M., Aerts C., Cuypers J., Vandenbussche B.,
Garrido R., Solano E.
<Astron. Astrophys., 475, 1159-1183 (2007)>
=2007A&A...475.1159D 2007A&A...475.1159D
ADC_Keywords: Stars, variable
Keywords: stars: variables: general - stars: binaries: general -
techniques: photometric - methods: statistical -
methods: data analysis
Abstract:
The fast classification of new variable stars is an important step in
making them available for further research. Selection of science
targets from large databases is much more efficient if they have been
classified first. Defining the classes in terms of physical parameters
is also important to get an unbiased statistical view on the
variability mechanisms and the borders of instability strips. Our goal
is twofold: provide an overview of the stellar variability classes
that are presently known, in terms of some relevant stellar
parameters; use the class descriptions obtained as the basis for an
automated `supervised classification' of large databases. Such
automated classification will compare and assign new objects to a set
of pre-defined variability training classes. For every variability
class, a literature search was performed to find as many well-known
member stars as possible, or a considerable subset if too many were
present. Next, we searched on-line and private databases for their
light curves in the visible band and performed period analysis and
harmonic fitting. The derived light curve parameters are used to
describe the classes and define the training classifiers. We compared
the performance of different classifiers in terms of percentage of
correct identification, of confusion among classes and of computation
time. We describe how well the classes can be separated using the
proposed set of parameters and how future improvements can be made,
based on new large databases such as the light curves to be assembled
by the CoRoT and Kepler space missions. The derived classifiers'
performances are so good in terms of success rate and computational
speed that we will evaluate them in practice from the application of
our methodology to a large subset of variable stars in the OGLE
database and from comparison of the results with published OGLE
variable star classifications based on human intervention. These
results will be published in a subsequent paper.
Description:
Light curve parameters for 1777 variable stars, belonging to 35 well
known stellar variability classes, are presented. These parameters are
used to define the classes and to construct the supervised classifiers
described in the paper. They serve as a direct input when using the
Gaussian mixture classification code, which is provided as well.
Instructions on how to run the code are given in the separate file
'gmreadme.txt'.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
list.dat 174 1642 List of stars used
defatts.dat 443 1777 Light curve parameters
codes.dat 49 35 Classification codes
gmreadme.txt 504 183 Readme file for the gmclass code
init.dat 24 69 Initialization file for the gmclass code
gmclass.f 225 573 gmclass classification code (FORTRAN)
--------------------------------------------------------------------------------
Byte-by-byte Description of file: list.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 22 A22 --- Name Star name
24- 25 I2 h RAh ? Simbad right ascension (J2000)
27- 28 I2 min RAm ? Simbad right ascension (J2000)
30- 36 F7.4 s RAs ? Simbad right ascension (J2000)
38 A1 --- DE- Simbad Declination sign (J2000)
39- 40 I2 deg DEd ? Simbad Declination (J2000)
42- 43 I2 arcmin DEm ? Simbad Declination (J2000)
45- 50 F6.3 arcsec DEs ? Simbad Declination (J2000)
52- 57 A6 --- var Variability code, in codes.dat file
58- 63 A6 --- var2 Second variability code, in codes.dat file
64- 90 A27 --- FileName Name of the file containing the lightcurve
92-118 A27 --- FileName2 Name of the file containing the lightcurve
120-146 A27 --- FileName3 Name of the file containing the lightcurve
148-174 A27 --- FileName4 Name of the file containing the lightcurve
--------------------------------------------------------------------------------
Byte-by-byte Description of file: defatts.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 27 A27 --- FileName Name of the file containing the lightcurve (1)
32- 44 F13.8 1/d f1 1th frequency
46- 58 F13.8 1/d f2 2th frequency
60- 72 F13.8 1/d f3 3th frequency
74- 86 F13.8 mag amp11 1th harmonic amplitude of f1
88-100 F13.8 mag amp12 2th harmonic amplitude of f1
102-114 F13.8 mag amp13 3th harmonic amplitude of f1
116-128 F13.8 mag amp14 4th harmonic amplitude of f1
130-142 F13.8 mag amp21 1th harmonic amplitude of f2
144-156 F13.8 mag amp22 2th harmonic amplitude of f2
158-170 F13.8 mag amp23 3th harmonic amplitude of f2
172-184 F13.8 mag amp24 4th harmonic amplitude of f2
186-198 F13.8 mag amp31 1th harmonic amplitude of f3
200-212 F13.8 mag amp32 2th harmonic amplitude of f3
214-226 F13.8 mag amp33 3th harmonic amplitude of f3
228-240 F13.8 mag amp34 4th harmonic amplitude of f3
242-254 F13.8 rad phi12 phase of amp12, if phase of amp11=0
256-268 F13.8 rad phi13 phase of amp13, if phase of amp11=0
270-282 F13.8 rad phi14 phase of amp14, if phase of amp11=0
284-296 F13.8 rad phi21 phase of amp21, if phase of amp11=0
298-310 F13.8 rad phi22 phase of amp22, if phase of amp11=0
312-324 F13.8 rad phi23 phase of amp23, if phase of amp11=0
326-338 F13.8 rad phi24 phase of amp24, if phase of amp11=0
340-352 F13.8 rad phi31 phase of amp31, if phase of amp11=0
354-366 F13.8 rad phi32 phase of amp32, if phase of amp11=0
368-380 F13.8 rad phi33 phase of amp33, if phase of amp11=0
382-394 F13.8 rad phi34 phase of amp34, if phase of amp11=0
396-408 F13.8 mag/d trend Slope of linear trend
410-422 F13.8 --- varrat Variance ratio (2)
424-436 F13.8 --- varred Variance reduction (3)
438-443 A6 --- var Variability code, in codes.dat file
--------------------------------------------------------------------------------
Note (1): The original light curves for these objects will be made
available soon, together with an explanation on the identifiers.
Note (2): Ratio of the variance after, to the variance before subtraction
of a least-squares fit with the dominant frequency 'f1' and its 4
harmonics (values between 0 and 1).
Note (3): Final variance reduction due to subtraction of all the periodic
signals (values close to 1 if the fit is good, close to 0 if the fit
is poor)
--------------------------------------------------------------------------------
Byte-by-byte Description of file: codes.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 5 A5 --- Code Code
9- 45 A37 --- Class Definition of the class
47- 49 I3 --- Ncl The total number of light curves present in
the dataset for the class
--------------------------------------------------------------------------------
Acknowledgements:
Jonas Debosscher, jonas(at)ster.kuleuven.be
Historical notes:
* 20-Nov-2007: Initial version
* 31-Mar-2008: New version of list.dat (Variability code corrected)
(End) Patricia Vannier [CDS] 19-Nov-2007