J/A+A/657/A138 Classification of Swift and XMM-Newton sources (Tranin+, 2022)
Probabilistic classification of X-ray sources applied to Swift-XRT and
XMM-Newton catalogs.
Tranin H., Godet O., Webb N. Primorac D.
<Astron. Astrophys. 657, A138 (2022)>
=2022A&A...657A.138T 2022A&A...657A.138T (SIMBAD/NED BibCode)
ADC_Keywords: Active gal. nuclei ; Binaries, X-ray ; X-ray sources
Keywords: catalogs - X-rays: general - X-rays: binaries - X-rays: galaxies -
methods: statistical
Abstract:
With the advent of very large X-ray surveys, an automated
classification of X-ray sources becomes increasingly valuable. This
work proposes a revisited naive Bayes classification of the X-ray
sources in the Swift-XRT and XMM- Newton catalogs into four classes -
AGN, stars, X-ray binaries (XRBs), and cataclysmic variables (CVs) -
based on their spatial, spectral, and timing properties and their
multiwavelength counterparts. An outlier measure is used to identify
objects of other natures. The classifier is optimized to maximize the
classification performance of a chosen class (here XRBs), and it is
adapted to data mining purposes. We augmented the X-ray catalogs with
multiwavelength data, source class, and variability properties. We
then built a reference sample of about 25000 X-ray sources of known
nature. From this sample, the distribution of each property was
carefully estimated and taken as reference to assign probabilities of
belonging to each class. The classification was then performed on the
whole catalog, combining the information from each property. Using the
algorithm on the Swift reference sample, we retrieved 99%, 98%, 92%,
and 34% of AGN, stars, XRBs, and CVs, respectively, and the false
positive rates are 3%, 1%, 9%, and 15%. Similar results are obtained
on XMM sources. When applied to a carefully selected test sample,
representing 55% of the X-ray catalog, the classification gives
consistent results in terms of distributions of source properties. A
substantial fraction of sources not belonging to any class is
efficiently retrieved using the outlier measure, as well as AGN and
stars with properties deviating from the bulk of their class. Our
algorithm is then compared to a random forest method; the two showed
similar performances, but the algorithm presented in this paper
improved insight into the grounds of each classification. This robust
classification method can be tailored to include additional or
different source classes and can be applied to other X-ray catalogs.
The transparency of the classification compared to other methods makes
it a useful tool in the search for homogeneous populations or rare
source types, including multi-messenger events. Such a tool will be
increasingly valuable with the development of surveys of unprecedented
size, such as LSST, SKA, and Athena, and the search for counterparts
of multi-messenger events.
Description:
The following tables represent 72% and 84% of the Swift and XMM-Newton
catalogs, respectively, taken as input of the classifier. They contain
all information used in the classification, as well as the output
results.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table6.dat 495 148438 Classification of 2SXPS sources
table7.dat 484 482771 Classification of 4XMM-DR10 sources
--------------------------------------------------------------------------------
See also:
VII/281 : GLADE v2.3 catalog (Dalya+, 2018)
IX/58 : 2SXPS Swift X-ray telescope point source catalogue (Evans+, 2020)
IX/59 : XMM-Newton Serendipitous Source Catalogue 4XMM-DR9 (Webb+, 2020)
Byte-by-byte Description of file: table6.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 22 A22 --- 2SXPS IAUNAME in 2SXPS
24- 32 F9.5 deg RAdeg Right ascension (J2000) of the X-ray
source, from 2SXPS
34- 42 F9.5 deg DEdeg Declination (J2000) of the X-ray source,
from 2SXPS
44- 47 F4.2 arcsec Err1sigma X-ray 1 sigma positional error from 2SXPS
49 I1 --- DetFlag [0/1] Detection flag, from 2SXPS
51 I1 --- FieldFlag [0/1] Field flad, from 2SXPS
53- 58 F6.2 --- SNR ? Signal-to-noise ratio,
from 2SXPS flux and error
60- 68 E9.3 mW/m2 Flux ? The mean X-ray observed flux
(0.3-10keV, powerlaw fit) from 2SXPS
70- 72 I3 --- Ndet Number of X-ray detections of the source
from XMM, Swift and Chandra
74 I1 --- Quality Number of quality rules followed by the
source (≥2 : Test Sample)
76- 84 F9.5 deg GLAT Galactic latitude
86- 93 F8.3 mas/yr pmGAIA ? Totla proper motion of the best GAIA
counterpart, if any
95-103 E9.3 10-7W LXGAIA ? Luminosity of the GAIA counterpart using
its parallax-based distance
105-113 F9.5 deg RAGdeg ? Right ascension (J2000) of the presumed
host center, from GLADE
115-123 F9.5 deg DEGdeg ? Declination (J2000) of the presumed
host center, from GLADE
125-133 F9.4 Mpc DistGLADE ? Luminosity distance of the presumed
host, from GLADE
135-140 F6.4 --- SepGLADE ? Angular separation to the host center,
relative to its semi-major axis
142-150 E9.3 10-7W LXGLADE ? Mean X-ray luminosity, computed using
the host distance
152-157 F6.3 --- HR1 ? First hardness ratio, from 2SXPS
159-164 F6.3 --- HR2 ? Second hardness ratio, from 2SXPS
166-173 F8.4 --- FitPowGamma ? Powerlaw photon index of the best
powerlaw spectral fit, from 2SXPS
175-181 F7.4 --- IntPowGamma ? Powerlaw photon index interpolated from
hardness ratios, from 2SXPS
183-189 F7.4 keV FitAPECkT ? Temperature APEC parameter of the best
APEC fit, from 2SXPS
191-197 F7.4 keV IntAPECkT ? Temperature parameter interpolated from
hardness ratios, from 2SXPS
199-206 F8.6 --- p-any-Opt ? Probability that the source have an
optical counterpart, from NWAY
208-212 F5.2 arcsec sepOpt ? Angular separation to the optical
counterpart
214-222 A9 --- RefOpt Origin of the optical counterpart
224-228 F5.2 [-] logFX/Fb ? Log. of the X-ray to b-band flux ratio,
from optical counterpart
230-234 F5.2 [-] logFX/Fr ? Log. of the X-ray to r-band flux ratio,
from optical counterpart
236-243 F8.6 --- p-any-IR ? Probability that the source have an
infrared counterpart, from NWAY
245-249 F5.2 arcsec sepIR ? Angular separation to the infrared
counterpart
251-257 A7 --- RefIR Origin of the infrared counterpart
259-264 F6.2 [-] logFX/FW1 ? Log. of the X-ray to W1 (or K) flux
ratio, from IR counterpart
266-271 F6.2 [-] logFX/FW2 ? Log. of the X-ray to W2 (or K) flux
ratio, from IR counterpart
273-277 F5.2 [-] logFratioSnap ? Log of max-to-med ratio from Swift
snapshots in most variable band
279-282 F4.2 [-] logFratioObs ? Log. of the max-to-min flux ratio,
from Swift/XMM/Chandra detections
284-291 F8.6 --- PbaC0 Posterior probability that the source
is an AGN
293-300 F8.6 --- PbaC1 Posterior probability that the source
is a star
302-309 F8.6 --- PbaC2 Posterior probability that the source
is an XRB
311-318 F8.6 --- PbaC3 Posterior probability that the source
is a CV
320-322 I3 --- classin ? Input class, for sources in the
reference sample
324 I1 --- classout Output class, given by the classification
326-344 A19 --- AltClass Alternative classifications (0-3)
if a property category is ignored
346-351 F6.3 --- outlier Outlier measure
353-360 F8.6 --- PbaC0loc Combined likelihood of location properties
for the class AGN
362-369 F8.6 --- PbaC1loc Combined likelihood of location properties
for the class star
371-378 F8.6 --- PbaC2loc Combined likelihood of location properties
for the class XRB
380-387 F8.6 --- PbaC3loc Combined likelihood of location properties
for the class CV
389-396 F8.6 --- PbaC0spe Combined likelihood of spectrum properties
for the class AGN
398-405 F8.6 --- PbaC1spe Combined likelihood of spectrum properties
for the class star
407-414 F8.6 --- PbaC2spe Combined likelihood of spectrum properties
for the class XRB
416-423 F8.6 --- PbaC3spe Combined likelihood of spectrum properties
for the class CV
425-432 F8.6 --- PbaC0mul Combined likelihood of multiwavelength
properties for the class AGN
434-441 F8.6 --- PbaC1mul Combined likelihood of multiwavelength
properties for the class star
443-450 F8.6 --- PbaC2mul Combined likelihood of multiwavelength
properties for the class XRB
452-459 F8.6 --- PbaC3mul Combined likelihood of multiwavelength
properties for the class CV
461-468 F8.6 --- PbaC0var Combined likelihood of variability
properties for the class AGN
470-477 F8.6 --- PbaC1var Combined likelihood of variability
properties for the class star
479-486 F8.6 --- PbaC2var Combined likelihood of variability
properties for the class XRB
488-495 F8.6 --- PbaC3var Combined likelihood of variability
properties for the class CV
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table7.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 21 A21 --- 4XMM IAUNAME in 4XMM-DR10
23- 31 F9.5 deg RAdeg Right ascension (J2000) of the X-ray
source, from 4XMM-DR10
33- 41 F9.5 deg DEdeg Declination (J2000) of the X-ray source,
from 4XMM-DR10
43- 46 F4.2 arcsec Err1sigma X-ray 1 sigma positional error from
4XMM-DR10
48 I1 --- SC-SUM-FLAG Summary flag, from 4XMM-DR10
50- 59 F10.1 --- SC-DET-ML Detection maximum likelihood,
from 4XMM-DR10
61- 71 E11.5 mW/m2 Flux The mean X-ray observed flux (0.2-12keV)
from 4XMM-DR10
73- 75 I3 --- Ndet Number of X-ray detections of the source
from XMM, Swift and Chandra
77 I1 --- Quality Number of quality rules followed by the
source (≥2 : Test Sample)
79- 87 F9.5 deg GLAT Galactic latitude
89- 96 F8.3 mas/yr pmGAIA ? Total proper motion of the best GAIA
counterpart, if any
98-106 E9.3 10-7W LXGAIA ? Luminosity of the GAIA counterpart using
its parallax-based distance
108-116 F9.5 deg RAGdeg ? Right ascension (J2000) of the presumed
host center, from GLADE
118-126 F9.5 deg DEGdeg ? Declination (J2000) of the presumed
host center, from GLADE
128-137 F10.4 Mpc DistGLADE ? Luminosity distance of the presumed
host of the source, from GLADE
139-144 F6.4 --- SepGLADE ? Angular separation to the host center,
relative to its semi-major axis
146-154 E9.3 10-7W LXGLADE ? Mean X-ray luminosity, computed using the
host distance
156-162 F7.4 --- SC-HR1 ? 1st hardness ratio, from 4XMM-DR10
164-170 F7.4 --- SC-HR2 ? 2nd hardness ratio, from 4XMM-DR10
172-178 F7.4 --- SC-HR3 ? 3rd hardness ratio, from 4XMM-DR10
180-186 F7.4 --- SC-HR4 ? 4rth hardness ratio, from 4XMM-DR10
188-193 F6.4 --- WAPO-GAMMA ? Photon index of the best powerlaw
spectral fit, from XMMFITCAT
195-202 F8.6 --- p-any-Opt ? Probability that the source have an
optical counterpart, from NWAY
204-207 F4.2 arcsec sepOpt ? Angular separation to the optical
counterpart
209-217 A9 --- RefOpt Origin of the optical counterpart
219-223 F5.2 [-] logFX/Fb ? Log. of the X-ray to b-band flux
ratio, from optical counterpart
225-229 F5.2 [-] logFX/Fr ? Log. of the X-ray to r-band flux
ratio, from optical counterpart
231-238 F8.6 --- p-any-IR ? Probability that the source have an
infrared counterpart, from NWAY
240-243 F4.2 arcsec sepIR ? Angular separation to the infrared
counterpart
245-251 A7 --- RefIR Origin of the infrared counterpart
253-257 F5.2 [-] logFX/FW1 ? Log. of the X-ray to W1 (or K) flux
ratio, from infrared counterpart
259-263 F5.2 [-] logFX/FW2 ? Log. of the X-ray to W2 (or K) flux
ratio, from infrared counterpart
265-268 F4.2 [-] logFratioObs ? Log. of the max-to-min flux ratio,
from Swift/XMM/Chandra detections
270-277 F8.6 --- PbaC0 Posterior probability that the source
is an AGN
279-286 F8.6 --- PbaC1 Posterior probability that the source
is a star
288-295 F8.6 --- PbaC2 Posterior probability that the source
is an XRB
297-304 F8.6 --- PbaC3 Posterior probability that the source
is a CV
306-311 I6 --- classin ? Input class, for sources in the
reference sample
313 I1 --- classout Output class, given by the classification
315-333 A19 --- AltClass Alternative classifications (0-3) if a
property category is ignored
335-340 F6.3 --- outlier Outlier measure
342-349 F8.6 --- PbaC0loc Combined likelihood of location properties
for the class AGN
351-358 F8.6 --- PbaC1loc Combined likelihood of location properties
for the class star
360-367 F8.6 --- PbaC2loc Combined likelihood of location properties
for the class XRB
369-376 F8.6 --- PbaC3loc Combined likelihood of location properties
for the class CV
378-385 F8.6 --- PbaC0spe Combined likelihood of spectrum properties
for the class AGN
387-394 F8.6 --- PbaC1spe Combined likelihood of spectrum properties
for the class star
396-403 F8.6 --- PbaC2spe Combined likelihood of spectrum properties
for the class XRB
405-412 F8.6 --- PbaC3spe Combined likelihood of spectrum properties
for the class CV
414-421 F8.6 --- PbaC0mul Combined likelihood of multiwavelength
properties for the class AGN
423-430 F8.6 --- PbaC1mul Combined likelihood of multiwavelength
properties for the class star
432-439 F8.6 --- PbaC2mul Combined likelihood of multiwavelength
properties for the class XRB
441-448 F8.6 --- PbaC3mul Combined likelihood of multiwavelength
properties for the class CV
450-457 F8.6 --- PbaC0var Combined likelihood of variability
properties for the class AGN
459-466 F8.6 --- PbaC1var Combined likelihood of variability
properties for the class star
468-475 F8.6 --- PbaC2var Combined likelihood of variability
properties for the class XRB
477-484 F8.6 --- PbaC3var Combined likelihood of variability
properties for the class CV
--------------------------------------------------------------------------------
Acknowledgements:
Hugo Tranin, hugo.tranin(at)irap.omp.eu
(End) Patricia Vannier [CDS] 08-Nov-2021