J/MNRAS/515/4711 Source classifier with WISE * PS1 3π DR2 (Beck+, 2022)
WISE-PS1-STRM neural network source classification and photometric redshifts
for WISE*PS1.
Beck R., Curt Dodds S., Szapudi I.
<Mon. Not. R. Astron. Soc. 515, 4711-4721 (2022)>
=2022MNRAS.515.4711B 2022MNRAS.515.4711B (SIMBAD/NED BibCode)
ADC_Keywords: Cross identifications ; Positional data ; Galaxies ; QSOs ;
Redshifts ; Photometry ; Optical ; Infrared
Keywords: methods: data analysis; methods: numerical; catalogues
large-scale structure of Universe
Abstract:
We cross-match between the WISE All-Sky and PS1 3π DR2 source
catalogues. The resulting catalogue has 354590570 objects,
significantly fewer than the parent PS1 catalogue, but its combination
of optical and infrared colours facilitate both better source
classification and photometric redshift estimation. We perform a
neural network-based classification of the objects into galaxies,
quasars, and stars, then run neural network-based photometric redshift
estimation for the galaxies. The star sample purity and quasar sample
completeness measures improve substantially, and the resulting
photo-z's are significantly more accurate in terms of statistical
scatter and bias than those calculated from PS1 properties alone. The
catalogue will be a basis for future large-scale structure studies,
and will be made available as a high-level science product via the
Mikulski Archive for Space Telescopes.
Description:
To extend PS1 into the infrared domain, we combine it with the WISE
All-Sky data set, and characterize the light sources by relying on
matched spectroscopic observations (mainly from the SDSS) as a
training set. This paper provides a detailed account of the WISE*PS1
Source Types and Redshifts with Machine learning catalogue
(WISE-PS1-STRM), a neural network source classification and photo-z
catalogue for the cross-match between WISE All-Sky and PS1 3π DR2.
This work significantly relies on Beck et al. (2021MNRAS.500.1633B 2021MNRAS.500.1633B),
which is a description of the PS1 3π DR1 photo-z and source
classification catalogue PS1-STRM. A sample of the full catalogue is
given in table.sam (i.e refer to
https://archive.stsci.edu/hlsp/wise-ps1-strm).
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
wiseps1.sam 323 1000 Neural network source class and photometric z
catalog from WISE All-Sky and PS1 3π DR2
--------------------------------------------------------------------------------
See also:
II/349 : The Pan-STARRS release 1 (PS1) Survey - DR1 (Chambers+, 2016)
J/ApJ/750/99 : The Pan-STARRS1 photometric system (Tonry+, 2012)
II/363 : The band-merged unWISE Catalog (Schlafly+, 2019)
J/AJ/132/926 : Galaxies in the Hubble Ultra Deep Field (Coe+, 2006)
II/311 : WISE All-Sky Data Release (Cutri+ 2012)
Byte-by-byte Description of file: wiseps1.sam
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 18 I18 --- PSI The main PS1 not unique source identifier
should be used to match with other PS1
tables (objID)
20- 33 F14.11 deg MeanRAdeg Right ascension from PS1 single-epoch
detections weighted mean at the mean epoch
given by epochMean (J2000) (raMean)
35- 47 F13.11 deg MeanDEdeg Declination from PS1 single-epoch
detections weighted mean at the mean epoch
given by epochMean (J2000) (decMean)
49- 59 F11.6 arcsec e_MeanRAdeg ?=-999 Right ascension standard deviation
from single-epoch detections (raMeanErr)
61- 71 F11.6 arcsec e_MeanDEdeg ?=-999 Declination standard deviation from
single-epoch detections (decMeanErr)
73- 87 E15.10 --- Sep The angular distance between the PS1 and
WISE detections (distance_Deg)
89-103 E15.10 arcsec2 sigma The combined PS1-WISE astrometric
uncertainty squared as used in section 3.1
(sqrErr_Arcsec)
105-119 E15.10 --- B The Bayes factor likelihood ratio of the
PS1-WISE cross-match as shown in equation
1 of the section 3.1 (BayesFactor)
121-138 I18 --- WISE Unique WISE identification number for the
Catalog source (cntr) (1)
140-153 F14.11 deg RAdeg Right ascension of WISE with respect to
the 2MASS PSC reference frame (J2000) (ra)
155-167 F13.11 deg DEdeg Declination of WISE with respect to the
2MASS PSC reference frame (J2000) (dec)
169-179 F11.9 arcsec e_RAdeg 1σ uncertainty in right ascension
coordinate (sigra)
181-191 F11.9 arcsec e_DEdeg 1σ uncertainty in declination
coordinate (sigdec)
193-204 F12.9 arcsec sigRADE The co-sigma of the equatorial position
uncertainties e_RAdeg and e_DEdeg
(sigradec) (2)
206-211 A6 --- Class The class assigned to the source (class)
(3)
213-223 F11.9 --- Galaxy The probability-like neural network output
for the galaxy class (prob_Galaxy)
225-235 F11.9 --- Star The probability-like neural network output
for the star class (prob_Star)
237-247 F11.9 --- QSO The probability-like neural network output
for the quasar class (prob_QSO)
249 I1 --- f_Class The extrapolation flag for the
classification (extrapolation_Class) (4)
251-255 I5 --- cellID The identifier of the nearest SOM cell in
the classification self organizing map SOM
as in section 3.4 (cellID_Class)
257-267 F11.6 --- zphot ?=-999 The Monte-Carlo photometric
redshift estimate z{phot} (z_phot) (5)
269-282 F14.9 --- e_zphot ?=-999 The calibrated redshift error
estimate calibrated with A=1.986 as
refered in section 3.5 and 4.2 (z_photErr)
284-297 F14.9 --- zphot0 The base photometric redshift estimate
z{phot,0} as refered in section 3.5 and
4.2 (z_phot0)
299-302 I4 --- f_zphot0 ?=-999 The extrapolation flag for the
photo-z estimation (extrapolation_Photoz)
(6)
304-317 F14.9 --- cellD ?=-999 The distance to the nearest SOM
cell centre in the photo-z self organizing
map denoted d{SOM} as in section 3.4
(cellDistance_Photoz)
319-323 I5 --- cellIDz ?=-999 The identifier of the nearest SOM
cell in the photo-z self organizing map
denoted as in section 3.4 (cellID_Photoz)
--------------------------------------------------------------------------------
Note (1): Refer to Cutri et al. 2014yCat.2328....0C 2014yCat.2328....0C, Cat. II/328 (https://
wise2.ipac.caltech.edu/docs/release/allsky/expsup/sec2_2a.html)
for more information.
Note (2): The co-sigma is related to the covariance between the position errors.
This maintains the sign information for the correlation. It is more
natural to carry the co-sigma along with the other uncertainties
instead of the covariance because the former is in the same units
as the other uncertainties.
Note (3): Using bD = 0.65 as exposed in section 4.1, classes are as :
"GALAXY", "STAR", "QSO" or "UNSURE" non-extrapolated and 1 if
extrapolated.
Note (4): Flag takes 0 if non-extrapolated and 1 if extrapolated (i.e see
Definition of Euclidean distance d{SOM} > 4.233 in section 3.4).
Note (5): Slightly less accurate than z{phot,0} as in sections 3.5 and 4.2.
Note (6): Flag takes 0 if non-extrapolated and 1 if extrapolated (i.e see
Definition of Euclidean distance d_{SOM} > 3.702 in section 3.4).
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Luc Trabelsi [CDS] 24-Sep-2025