J/A+A/710/A191 MGQPC photo-z and high-probability pair catalogs (Zhu+, 2026)
Search for quasar pairs with Gaia astrometric data. II.
Photometric redshift prediction with machine learning for the MGQPC catalogue.
Zhu X., Chen Q., Jing L., Deng Z., Xia J.-Q., Zhang Y., Wu J.
<Astron. Astrophys. 710, A191 (2026)>
=2026A&A...710A.191Z 2026A&A...710A.191Z (SIMBAD/NED BibCode)
ADC_Keywords: QSOs ; Redshifts ; Optical
Keywords: methods: data analysis - methods: statistical -
techniques: photometric - catalogs - galaxies: active -
quasars: general
Abstract:
The identification of physically associated kiloparsec-scale quasar
pairs is important for understanding galaxy evolution, the growth of
supermassive black holes, and the co-evolution of supermassive black
holes and their host galaxies. However, their intrinsic rarity and
high contamination from stellar superpositions and projected
alignments in photometric catalogues necessitate efficient selection
methods. This challenge is amplified as wide-field imaging surveys
such as the Legacy Survey of Space and Time (LSST) and Euclid will
discover vast numbers of quasar candidates, most lacking immediate
spectroscopic follow-up, making accurate photometric redshifts
(photo-z) essential for pre-selection.
This work aims to develop a robust machine-learning framework to
produce accurate photo-z point estimates and well-calibrated
probability density functions (PDFs) for quasars. The primary
application is to systematically identify high-probability quasar pair
candidates within the MGQPC catalogue, where MGQPC refers to Quasar
Pair Candidates derived from the cross-match of MQC and Gaia, by
assessing the redshift consistency of pair components.
We constructed two large, spectroscopically confirmed quasar samples
with multi-wavelength photometry: an SDSS-based sample (KSTS) and a
DESI-LS-based sample (KSTD). We employed the CatBoost algorithm for
point-estimate photo-z regression and the FlexZBoost framework for
full photo-z PDF estimation. A feature selection process guided by
SHAP (SHapley Additive exPlanations) was implemented to identify the
most predictive photometric features, including colours and magnitudes
from optical bands, such as u, g, r, i, z, G, GBP, and GRP, and
infrared bands, such as W1 and W2.
Our photo-z workflow achieved a robust performance, with a normalised
median absolute deviation of sigmaNMAD=0.036 and an outlier fraction
of 5.6% on the test sample. The resulting photo-z PDFs were well
calibrated, as verified by Probability Integral Transform (PIT)
diagnostics. Control tests indicate that the better performance of the
SDSS-based sample is primarily attributable to the inclusion of the u
band, although survey-dependent photometric differences in LS10,
including bright-source systematics, may also play a contributing
role. Applying this trained model to the MGQPC catalogue, we
identified 185 high-probability quasar pair candidates based on the
consistency between their photo-z estimates. This candidate list
included 20 systems subsequently confirmed as genuine physical pairs
by independent spectroscopic observations.
We demonstrate that modern machine-learning techniques, combining
gradient-boosted trees for point estimates and conditional density
estimation for PDFs, can deliver precise and reliable photo-zs for
quasars. This enables the efficient filtering and prioritisation of
rare quasar pairs from large photometric catalogues. The resulting
photo-z catalogue for the MGQPC provides a valuable resource for
future spectroscopic follow-up campaigns to study dual supermassive
black holes and their co-evolution.
Description:
Table 6 provides the photometric-redshift catalogue associated with
the MGQPC systems. Each GroupID corresponds to one candidate system
and occupies two consecutive rows, ordered as Member A followed by
Member B. The coordinates are taken from the updated MGQPC master
table used in this work. When both KSTS and KSTD solutions are
available for the same source, the KSTS solution is adopted; otherwise
the KSTD solution is used. Blank fields indicate that no valid photo-z
solution is available for that source in the current prediction files.
Table 7 lists the 185 high-probability quasar-pair candidates. Each
GroupID corresponds to one system and occupies two consecutive rows,
ordered as Member A followed by Member B.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table6.dat 132 8224 Photometric redshift catalogue
table7.dat 109 370 High-probability pair candidates
--------------------------------------------------------------------------------
See also:
J/A+A/707/A30 : Qasar pairs with Gaia astrometric data (Chen+, 2026)
Byte-by-byte Description of file: table6.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 4 I4 --- GroupID Identifier of the candidate system
6- 17 F12.8 deg RAdeg Gaia DR3 right ascension (ICRS) at Ep=2016.0
19- 30 F12.8 deg DEdeg Gaia DR3 declination (ICRS) at Ep=2016.0
32- 41 F10.7 --- zCat ? CatBoost photo-z (1)
43- 52 F10.7 --- zMean ? Mean redshift (2)
54- 63 F10.7 --- zMed ? Median redshift (2)
65- 74 F10.7 --- zPeak ? Peak redshift (2)
76- 85 F10.7 --- e_zPeak ? Uncertainty on peak redshift (2)
87- 88 I2 --- Npeak ? Number of peak used for zPeak (2)
90- 99 F10.7 --- zL68 ? 68% confidence level redshift lower value (2)
101-110 F10.7 --- zU68 ? 68% confidence level redshift upper value (2)
112-121 F10.7 --- e_zU68 ? Uncertainty on zU68 (2)
123-132 F10.7 --- e_zL68 ? Uncertainty on zL68 (2)
--------------------------------------------------------------------------------
Note (1): blank if no valid.
Note (2): blank if no valid PDF solution.
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table7.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 4 I4 --- GroupID Identifier of the candidate system
6- 27 A22 --- SysName Pair name designated by the center coordinates
29- 40 F12.8 deg RAdeg Gaia DR3 right ascension (ICRS) at Ep=2016.0
42- 53 F12.8 deg DEdeg Gaia DR3 declination (ICRS) at Ep=2016.0
55- 63 F9.6 mag Gmag Gaia G-band magnitude
65- 74 F10.7 --- z Adopted redshift
76- 85 F10.6 arcsec Sep Angular separation
87- 96 F10.6 kpc rp Projected separation
98-109 F12.6 km/s Dvr Line-of-sight velocity difference
--------------------------------------------------------------------------------
Acknowledgements:
From Xingyu Zhu, zhuxingyu(at)bnu.edu.cn
The authors thank the anonymous referee for the valuable comments that
improved the quality and clarity of the manuscript. This work has been
supported by the National Key R&D Program of China (2025YFA1614101,
2021YFA0718500) and by the Chinese National Natural Science Foundation
grant No. 12333001. J.-Q. Xia acknowledges support from the China
Manned Space Program (grant Nos. CMS-CSST-2025-A01 and
CMS-CSST-2025-A04).
We acknowledge the use of public data from the following facilities:
Gaia, SDSS, LAMOST, DESI, and DESI Legacy Surveys. This research made
use of the following software packages: Astropy (Astropy Collaboration
et al. 2013, 2018, 2022), Matplotlib (Hunter 2007), NumPy (van der
Walt et al. 2011; Harris et al. 2020), Pandas (Wes McKinney 2010;
pandas development team 2022), TOPCAT (Taylor 2005), HumVI (Marshall
et al. 2015, 2016).
References:
Chen et al., Paper I, 2026A&A...707A..30C 2026A&A...707A..30C, Cat. J/A+A/707/A30
License: CC-BY-4.0
(End) Patricia Vannier [CDS] 14-Apr-2026