Pith. sign in

REVIEW 4 major objections 5 minor 64 references

Photometric metallicities of 0.8 million KiDS stars

T0 review · 4 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Random forest models trained on spectra recover photometric metallicities for 820,055 KiDS/VIKING stars with about 0.28 dex scatter over -2 < [Fe/H] < 0.5.

desk verdict Useful new catalog with honestly reported external validation; the headline metallicity scatter holds up, but the uncertainty budget doesn't propagate to the final catalog's wider selection cuts. read the letter →

arxiv 2502.06512 v1 pith:GTSESLF6 submitted 2025-02-10 astro-ph.GA astro-ph.SR

Bao-Kun Sun , Bing-Qiu Chen , Xiao-Wei Liu This is my paper · ORCID
classification astro-ph.GAastro-ph.SR
keywords photometricmetallicityrandomforestKiDSsurveyVIKINGstellareffectivetemperatureabsolutemagnitudeMilkyWaystructureGalacticarchaeology
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims that the metal content of a star can be read from nine broadband colors alone, without a spectrum. The authors train a random forest on spectroscopic labels from LAMOST, SEGUE, APOGEE, GALAH, plus a curated very-metal-poor sample, and apply it to KiDS/VIKING photometry to estimate $[\mathrm{Fe/H}]$, effective temperature, and $g$-band absolute magnitude for 820,055 stars. Against held-out spectroscopic data the photometric metallicities scatter by about 0.28 dex in the range $-2 < [\mathrm{Fe/H}] < 0.5$, with comparable agreement against several independent catalogs. If the method holds, this catalog extends chemical-abundance mapping to the much larger and fainter photometric sky, including distant halo stars that spectroscopy cannot easily reach.

What carries the argument

The engine is a two-stage random forest pipeline. A random forest classifier first separates dwarfs from giants using eight dereddened KiDS/VIKING colors plus nine dereddened magnitudes, after an empirical giant polygon on the Kiel diagram defines the training labels. Then separate random forest regressors for dwarfs and giants map the eight intrinsic colors to $[\mathrm{Fe/H}]$, $T_{\mathrm{eff}}$, and $M_g$. A grid-averaging step rebalances the spectroscopically biased training sample, and an ensemble of 28 models (seven random data splits times four random seeds) provides internal uncertainty; total uncertainty is obtained by adding the label-prediction dispersion. The eight colors are the load-bearing inputs, so the method's sensitivity to metallicity comes almost entirely from the $u$-band and the optical-to-infrared color baselines.

What would settle it

Cross-match the released catalog against a large independent spectroscopic sample in the same sky area whose stars were not used in training, and compute the residual $[\mathrm{Fe/H}]_{\rm phot} - [\mathrm{Fe/H}]_{\rm spec}$ in bins of true metallicity and color. The claim of about 0.28 dex precision is falsified if the residual root-mean-square in the range $-2 < [\mathrm{Fe/H}] < 0.5$ substantially exceeds 0.3 dex, or if the mean residual varies systematically with $(g-r)_0$ by more than about 0.1 dex.

Watch

Extended reading notes

Core claim

The central claim is that a data-driven mapping from dereddened colors to stellar parameters generalizes across the KiDS/VIKING footprint. Specifically, the authors report that random forest regressors, trained after a dwarf/giant classification step, recover spectroscopic $[\mathrm{Fe/H}]$ with a mean offset of 0.00 dex and a scatter of 0.28 dex for stars with $-2 < [\mathrm{Fe/H}] < 0.5$, effective temperature within 149 K, and absolute $g$-band magnitude within 0.36 mag; below $[\mathrm{Fe/H}] \approx -2$ the model systematically overestimates metallicities and is not claimed reliable. The trained models were applied to 820,055 stars, yielding a public catalog of 814,383 dwarfs and 5,672 giants with positions, metallicities, temperatures, absolute magnitudes, distances, and per-star uncertainties. Validation against independent measurements (Pristine, DESI-based analyses, SMSS-based analyses, Gaia XP temperatures, cluster distances) shows small mean offsets and comparable scatter, which the paper interprets as evidence that the photometric estimates are physically meaningful rather than artifacts of the training labels.

Load-bearing premise

The central assumption is that the nine dereddened KiDS/VIKING colors of a star encode enough metallicity information that a model trained on a few thousand spectroscopic stars, plus curated metal-poor stars, continues to be accurate for the 0.8 million stars it was not trained on.

Editorial extensions

If this is right

  • The released catalog gives about 0.8 million stars with photometric $[\mathrm{Fe/H}]$, $T_{\mathrm{eff}}$, $M_g$, and distances, enough to trace metallicity gradients, halo substructure, and the thin/thick disk separation in the KiDS footprint.
  • At the quoted 0.28 dex scatter in the $-2 < [\mathrm{Fe/H}] < 0.5$ range, the sample can be split into metal-poor, intermediate, and metal-rich populations for statistical chemical-cartography studies.
  • The method reaches stars fainter than spectroscopic surveys typically measure; the paper notes the giant star subsample can be traced beyond 40 kpc, into the outer halo.
  • The pipeline is designed to be reused: the authors state it will be applied to the Mephisto and CSST photometric surveys.
  • The giant classification is far less pure than the dwarf classification (31% purity on the test set), so giant-subsample results should be filtered by the published giant probability before use.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the color–metallicity mapping is as stable as the validation suggests, the same training procedure could be retargeted to other multi-band surveys with overlapping filter sets to build comparable metallicity maps on different footprints; this is an extension the paper does not test.
  • The drop in reliability below $[\mathrm{Fe/H}] \approx -2$ is likely a generic limitation of broadband photometric metallicities, not just this pipeline; a practical next step would be to train a separate classifier that flags very metal-poor candidates instead of regressing $[\mathrm{Fe/H}]$ continuously.
  • Because the catalog includes distances and positions, it could be cross-correlated with Gaia astrometry to measure kinematic–chemical signatures of the halo; the paper stops at presenting the catalog and its basic validation.
  • The ensemble-of-28 uncertainty scheme gives per-star error bars but assumes the training labels themselves are unbiased; a testable follow-up would compare the published per-star uncertainties against the scatter from repeated measurements in overlapping survey regions.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper presents a data-driven method to estimate stellar [Fe/H], effective temperature, and g-band absolute magnitude from nine-band KiDS/VIKING photometry, applying random forest classification and regression. The algorithm is trained on spectroscopic labels from LAMOST, SEGUE, APOGEE, GALAH, and a literature sample of very metal-poor stars, plus Gaia EDR3/Bailer-Jones et al. (2021) distances for absolute-magnitude calibration, after gridding the training samples to mitigate label imbalance. The resulting catalog contains 820,055 stars, of which 814,383 are classified as dwarfs and 5,672 as giants. The authors report internal test scatters of 0.28 dex for [Fe/H] in the range -2 to 0.5 dex, 149 K for Teff, and 0.36 mag for Mg, and they compare their results with independent catalogs (Zhang et al. 2024, Martin et al. 2023, Lu et al. 2024, Lin et al. 2022, Huang et al. 2022, Vallenari et al. 2023) as well as with cluster and parallax-based distances.

Significance. If the claimed accuracy extends to the released catalog, this is a useful resource for Galactic structure and chemical-evolution studies, providing photometric metallicities for roughly 0.8 million stars at high Galactic latitude. The paper has several genuine strengths: it uses a wide wavelength baseline (u through Ks), it augments the training set with very metal-poor stars, it attempts zeropoint corrections among spectroscopic surveys, it releases the catalog publicly, and it validates against multiple independent datasets. The external comparisons give roughly consistent dispersions (0.28-0.31 dex in [Fe/H]), which supports the central metallicity claim for the subset of stars covered by those comparisons. However, the headline accuracy is established on a clean, low-extinction, high-S/N training sample, while the final catalog relaxes the reddening and S/N cuts; the distance validation is partly circular; and the per-star uncertainties do not propagate photometric and reddening errors. These issues are load-bearing for the catalog's advertised accuracy and require additional analysis or careful reframing.

major comments (4)
  1. [§4.1 vs. §2.3 and §4.2] The accuracy claimed in the abstract and conclusion (0.28 dex for [Fe/H]) is demonstrated on stars selected with the training cuts of Section 2.3 (E(B-V)<0.1 mag, per-band photometric errors <0.08 mag, S/N>30), and the independent comparisons in Section 4.2 inherit a similar selection. The final catalog in Section 4.1 relaxes these cuts: it only requires photometric errors <0.1 mag and class_star>0.9, with no reddening or S/N cut. Because the RFR inputs are dereddened colors, errors in the Schlegel et al. (1998) E(B-V) propagate directly into [Fe/H], and the u-g color is among the most extinction-sensitive inputs. The paper never reports external validation residuals as a function of E(B-V), apparent magnitude, or photometric error, so the 0.28 dex scatter has not been shown to hold for the fainter or more reddened stars in the final catalog. Please provide such stratified validation or restrict the accuracy claim to the subset satisfying the training cuts.
  2. [§2.1, §2.3, §4.2 (Fig. 8)] The distance estimates are calibrated on Bailer-Jones et al. (2021) distances: Section 2.1 states that distances are taken from that catalog, and Section 2.3 constructs the KG training sample from Gaia EDR3 with tight parallax-quality cuts. The right panel of Fig. 8 then compares the resulting distances with Bailer-Jones et al. (2021) and reports agreement, which is a circular test and cannot by itself confirm the reliability of the distance estimates. The Blanco 1 comparison is an independent check, but it uses only 85 stars in one open cluster. Please validate the distances against an independent sample (e.g., asteroseismic distances, additional clusters, or Gaia parallaxes excluded from training) or soften the distance-validation claim accordingly.
  3. [§3.3 and §4.1] The per-star uncertainties e_feh, e_Teff, and e_Mg listed in the catalog are computed by combining the spread of the 28 model predictions with the dispersion of training residuals, as described in Section 3.3. The text itself notes that this 'does not account for the full predictive error, as it excludes uncertainties from the input measurements,' yet Table 3 presents these values as the uncertainties for each catalog star without further qualification. Please rename or redefine these columns as internal precision estimates and, where feasible, add a photometric and reddening error term; otherwise users may interpret e_feh as a total uncertainty that the paper does not actually claim to provide.
  4. [§2.3, §3.2, Fig. 2] The headline scatters of 0.28 dex, 149 K, and 0.36 mag are measured on a holdout from the same gridded training sample used to build the regression models. The grid averaging described in Section 2.3 smooths the labels, so this in-sample scatter is expected to be an optimistic estimate of accuracy on real stars. The external comparisons in Section 4.2 are the more meaningful accuracy check, and their agreement with the internal values (σ=0.29, 0.28, and 0.31 dex for the Zhang, Martin, and Lu comparisons) is encouraging. The paper should state explicitly that the internal test scatter is not an independent accuracy estimate and should present the external, stratified validation as the primary accuracy metric.
minor comments (5)
  1. [§2.3] The text first reports 9312 synthetic stars for the KS sample (8248 dwarfs plus 1064 giants) but later says '9,001 stars (KS sample)'; please correct this numerical inconsistency.
  2. [§3.3] The sentence 'The systematic errors are measured by the spread of predictions across the 28 RFR models' conflicts with the immediately preceding definition of internal uncertainty as the standard deviation across those same models; 'systematic' is defined two sentences later as the dispersion of true-minus-predicted labels in the training sample, so please reword for consistency.
  3. [Table 3] The column header 'Photonetric magnitudes' contains a typo and should read 'Photometric magnitudes'; also, the VIKING band labels should be uppercase Z/Y/J/H/Ks to match the survey notation used elsewhere in the text.
  4. [§4.2] For the right panel of Fig. 8, the text does not state whether the 99,526 common sources overlap with the KG training sample; please clarify this overlap, since the circularity concern depends on it.
  5. [§4.2] The statement that Lin et al. (2022) 'may have overestimated the metallicities of their stars' is presented without a detailed diagnostic beyond agreement with the other comparisons; please either provide a more quantitative assessment or soften this claim.

Circularity Check

1 steps flagged · score 4.0 of 10

Central metallicity and Teff results are independently benchmarked; the distance validation is partly circular because it reuses the Bailer-Jones catalog that generated the Mg training labels.

  1. fitted input called prediction [Section 2.1 and Section 4.2 (Fig. 8, right panel)]
    "Rather than directly inverting Gaia parallax measurements to estimate distances, we employed the distance catalog from Bailer-Jones et al. (2021), which provides distance estimates for nearly 1.47 billion stars. ... To assess the accuracy of the distances in our sample, we compared our results with two other independent datasets: ... distances derived from Gaia parallaxes by Bailer-Jones et al. (2021). ... Our distances are in good agreement with those from Bailer-Jones et al. (2021), with a mean difference of 0.048 kpc and a dispersion of 0.38 kpc."

    The KG training sample's absolute-magnitude labels are built from this same Bailer-Jones distance catalog, and the catalog distances in the final sample are derived from the photometrically predicted Mg. The right panel of Fig. 8 therefore compares the model output with the catalog that supplied the training labels, so the agreement is at least partly built in. The paper explicitly calls this an independent validation, but only the Cantat-Gaudin open-cluster comparison is external to the calibration chain.

full rationale

The main metallicity claim is a standard supervised regression: random forest models trained on LAMOST, SEGUE, APOGEE, GALAH, and literature metal-poor labels, with dereddened KiDS/VIKING colors as inputs, tested on a held-out split. This is not circular, and the claimed 0.28 dex scatter is independently corroborated by external comparisons with Zhang et al. (2024, sigma = 0.29 dex), Martin et al. (2023, sigma = 0.28 dex), and Lu et al. (2024, sigma = 0.31 dex). The Teff estimates also agree with independent catalogs such as Huang et al. (2022) and Gaia XP. The one genuine circularity is the distance validation: the paper adopts Bailer-Jones et al. (2021) as the distance reference, uses the resulting Mg values as the KG training target, and then presents agreement with that same catalog as an independent check. Because the distance estimate is ancillary to the paper's central metallicity claim and does have one external cluster check, the overall circularity score is modest rather than severe. The final catalog's relaxation of training cuts such as E(B-V) < 0.1 and S/N > 30 is a generalization concern, not a circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the accuracy and consistency of external spectroscopic labels, the reddening corrections, and the transferability of a random forest trained on a gridded subsample. No new physical entities are introduced.

free parameters (5)
  • Grid step sizes for training set averaging = KS: 150 K, 0.05 dex, 0.02 dex; KG: 0.01 mag
    The training sample is replaced by a grid in Teff, log g, and [Fe/H] (or Mg and (g-r)0). The bin widths are chosen by hand and affect the resolution and sample size of the training set (Section 2.3).
  • Random forest hyperparameters = not specified
    The number of trees, maximum depth, and feature sampling are not reported. These control the bias-variance tradeoff and are tuned by the authors (Section 3).
  • Zeropoint corrections for spectroscopic surveys = not specified
    The paper states that LAMOST, GALAH, and SEGUE metallicities are zeropoint-corrected using cross-matches with APOGEE, but the offset values are not listed, making the scale harmonization uncheckable (Section 2.2).
  • Selection thresholds = photometric error < 0.08 mag, E(B-V) < 0.1, S/N > 30, etc.
    These manual cuts define the training and final samples and directly influence the parameter coverage and the reported uncertainties (Sections 2.3 and 4.1).
  • Kiel diagram giant/dwarf classification boundary = empirical polygon, not quantified
    The boundary separating giants from dwarfs is drawn by eye in Figure 1. Classification errors propagate into the separate regression models, and the giant class has only 31% purity (Table 2).
assumptions (5)
  • domain assumption Spectroscopic labels from LAMOST, SEGUE, APOGEE, and GALAH are accurate and on a common scale after zeropoint corrections.
    These labels are used as ground truth for training. Any systematic offset in the labels becomes a systematic offset in the photometric estimates. Invoked in Sections 2.2 and 2.3.
  • domain assumption Reddening corrections from Schlegel et al. (1998) and extinction coefficients from Kuijken et al. (2019) are accurate for the high-latitude KiDS fields.
    Colors are dereddened before training; errors in E(B-V) directly bias photometric metallicities. Stated in Section 2.1.
  • domain assumption The nine-band photometry carries enough metallicity information to learn [Fe/H] over -2 to 0.5 dex.
    The method assumes a learnable mapping from colors to metallicity. The paper itself notes in Section 3.2 that the broad u-band is less sensitive below -2 dex, so the range of reliable extrapolation is limited.
  • ad hoc to paper Random forest regression trained on gridded synthetic examples generalizes to the full KiDS sample.
    The authors replace the measured training stars with 9,001 or 26,996 grid points and then train on those. This smoothing is specific to this paper and is not derived from data.
  • domain assumption The empirical Kiel-diagram giant/dwarf classification transfers to the final catalog.
    The giant class is recovered with 47% completeness and 31% purity in the test set (Table 2). Applying the classifier to 820,055 stars propagates this uncertainty into the giant-specific metallicity and Teff estimates.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Photometric metallicities of 0.8 million KiDS stars." pith.science (2026). https://pith.science/paper/GTSESLF6

@misc{pith2026250206512,
  author       = {Pith},
  title        = {Pith review of: Photometric metallicities of 0.8 million KiDS stars},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GTSESLF6}},
  note         = {Machine review of arXiv:2502.06512}
}
read the original abstract

Accurate determinations of metallicity for large, complete stellar samples are essential for advancing various studies of the Milky Way. In this paper, we present a data-driven algorithm that leverages photometric data from the KiDS and the VIKING surveys to estimate stellar absolute magnitude, effective temperature and metallicities. The algorithm is trained and validated using spectroscopic data from LAMOST, SEGUE, APOGEE, and GALAH, as well as a catalog of very metal-poor stars from the literature, and Gaia EDR3 data. This approach enables us to estimate metallicities, effective temperatures, and g-band absolute magnitudes for approximately 0.8 million stars in the KiDS dataset. The photometric metallicity estimates exhibit an uncertainty of around 0.28 dex when compared to spectroscopic studies, within the metallicity range of -2 dex to 0.5 dex. The photometric effective temperature estimates have an uncertainty of around 149 K, while the uncertainty in the absolute magnitude is approximately 0.36 mag. The metallicity estimates are reliable for values down to about -2 dex. This catalog represents a valuable resource for studying the structure and chemical properties of the Milky Way, offering an extensive dataset for future investigations into Galactic formation and evolution.

Figures

Figures reproduced from arXiv: 2502.06512 by the authors.

Figure 1
Figure 1. The Kiel diagram for the KS sample used in the dwarf-giant classification. The red polygon encloses stars classified as giants, while those outside are classified as dwarfs. 3 METHODS This section describes the methodology employed in this study. First, we distinguish between giant and dwarf stars using a Random Forest Classifier (RFC) (Breiman 2001). Once the stars are classified, we apply two separate Random Fores… view at source ↗
Figure 2
Figure 2. Comparisons between true and RFR-predicted [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Uncertainty estimates for Teff (left panel), [Fe/H] (middle panel) and Mg (right panel) as a func￾tion of the predicted values for the training samples. The red colormap represents the internal uncertainty, computed from the spread of predictions across the 28 RFR models. The green line shows the running dispersion derived from the training labels, representing systematic errors. The blue colormap depicts the 2D his… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Spatial number-density distributions for stars in the final sample. The panel shows the [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Comparison of our effective temperature estimates with those from Zhang et al. (2024, left panel), [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Comparison of our metallicity estimates with those from Zhang et al. (2024, left panel), Martin et al. [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Comparisons of our findings with those of Lin et al. (2022) in terms of [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Left panel: Histogram of our distance estimates for the member stars of the open cluster Blanco [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 61 canonical work pages

  1. [1]

    2018, arXiv preprint arXiv:1807.09409 10

    Abuter, R., Amorim, A., Anugu, N., et al. 2018, arXiv preprint arXiv:1807.09409 10

    arXiv 2018

  2. [2]

    S., Youakim, K., Gonz ´alez Hern ´andez, J

    Aguado, D. S., Youakim, K., Gonz ´alez Hern ´andez, J. I., et al. 2019, Monthly Notices of the Royal Astronomical Society, 490, 2241 4

    work page 2019

  3. [3]

    An, D., & Beers, T. C. 2020, The Astrophysical Journal, 897, 39 2

    work page 2020

  4. [4]

    2021, The Astronomical Journal, 161, 147 3, 12, 13, 14

    Bailer-Jones, C., Rybizki, J., Fouesneau, M., Demleitner, M., & Andrae, R. 2021, The Astronomical Journal, 161, 147 3, 12, 13, 14

    work page 2021

  5. [5]

    2001, Machine learning, 45, 5 6

    Breiman, L. 2001, Machine learning, 45, 5 6

    work page 2001

  6. [6]

    G., Vallenari, A., Prusti, T., et al

    Brown, A. G., Vallenari, A., Prusti, T., et al. 2021, Astronomy & Astrophysics, 649, A1 2

    work page 2021

  7. [7]

    2020, Astronomy & Astrophysics, 640, A1 12, 13

    Cantat-Gaudin, T., Anders, F., Castro-Ginard, A., et al. 2020, Astronomy & Astrophysics, 640, A1 12, 13

    work page 2020

  8. [8]

    2011, The Messenger, 146, 27 3 16 B.-K

    Capaccioli, M., & Schipani, P. 2011, The Messenger, 146, 27 3 16 B.-K. Sun et al

    work page 2011

Show all 64 references
  1. [9]

    2011, Astronomy & Astrophysics, 530, A138 1

    Casagrande, L., Sch¨onrich, R., Asplund, M., et al. 2011, Astronomy & Astrophysics, 530, A138 1

    2011

  2. [10]

    e., Moles, M., Crist´obal-Hornillos, D., et al

    Cenarro, A. e., Moles, M., Crist´obal-Hornillos, D., et al. 2019, Astronomy & Astrophysics, 622, A176 2

    2019

  3. [11]

    2018, Monthly Notices of the Royal Astronomical Society, 476, 3278 2

    Chen, B., Liu, X., Yuan, H., et al. 2018, Monthly Notices of the Royal Astronomical Society, 476, 3278 2

    2018

  4. [12]

    Collaboration, D., Aghamousa, A., Aguilar, J., et al. 2016 11

    2016

  5. [13]

    2024, The Astronomical Journal, 168, 33pp 11 Da Costa, G

    Collaboration, D., Adame, A., Aguilar, J., et al. 2024, The Astronomical Journal, 168, 33pp 11 Da Costa, G. S., Bessell, M. S., Mackey, A., et al. 2019, Monthly Notices of the Royal Astronomical Society, 489, 5900 4 De Silva, G. M., Freeman, K. C., Bland-Hawthorn, J., et al. 2...

    2024

  6. [14]

    2013, The Messenger, 154, 32 3

    Edge, A., Sutherland, W., Kuijken, K., et al. 2013, The Messenger, 154, 32 3

    2013

  7. [15]

    2023, The Astrophysical Journal Supplement Series, 268, 9 2 Fern´andez-Alvar, E., Kordopatis, G., Hill, V ., et al

    Fan, Z., Zhao, G., Wang, W., et al. 2023, The Astrophysical Journal Supplement Series, 268, 9 2 Fern´andez-Alvar, E., Kordopatis, G., Hill, V ., et al. 2021, Monthly Notices of the Royal Astronomical Society, 508, 1509 2

    2023

  8. [16]

    2016, Astronomy & Astrophysics, 595 2

    Gaia, C., Bono, G., et al. 2016, Astronomy & Astrophysics, 595 2

    2016

  9. [17]

    2018, Astronomy & Astrophysics, 616 2

    Gaia, C., Brown, A., Vallenari, A., et al. 2018, Astronomy & Astrophysics, 616 2

    2018

  10. [18]

    2023, arXiv preprint arXiv:2306.05086 9

    Guiglion, G., Nepal, S., Chiappini, C., et al. 2023, arXiv preprint arXiv:2306.05086 9

    2023 arXiv

  11. [19]

    2024, Astronomy & Astrophysics, 682, A9 2

    Guiglion, G., Nepal, S., Chiappini, C., et al. 2024, Astronomy & Astrophysics, 682, A9 2

    2024

  12. [20]

    2021, The Astrophysical Journal, 906, 47 7

    Guo, H.-L., Chen, B.-Q., Yuan, H.-B., et al. 2021, The Astrophysical Journal, 906, 47 7

    2021

  13. [21]

    2024, arXiv preprint arXiv:2405.18120 1

    Hackshaw, Z., Hawkins, K., Filion, C., et al. 2024, arXiv preprint arXiv:2405.18120 1

    2024 arXiv

  14. [22]

    H., & Friedman, J

    Hastie, T., Tibshirani, R., Friedman, J. H., & Friedman, J. H. 2009, The elements of statistical learning: data mining, inference, and prediction, V ol. 2 (Springer) 6

    2009

  15. [23]

    2024, arXiv preprint arXiv:2404.01269 2

    Hattori, K. 2024, arXiv preprint arXiv:2404.01269 2

    2024 arXiv

  16. [24]

    2019, The Astrophysical Journal Supplement Series, 243, 7 2

    Huang, Y ., Chen, B.-Q., Yuan, H.-B., et al. 2019, The Astrophysical Journal Supplement Series, 243, 7 2

    2019

  17. [25]

    C., Wolf, C., et al

    Huang, Y ., Beers, T. C., Wolf, C., et al. 2022, The Astrophysical Journal, 925, 164 2, 11, 13

    2022

  18. [26]

    C., Yuan, H., et al

    Huang, Y ., Beers, T. C., Yuan, H., et al. 2023, The Astrophysical Journal, 957, 65 2 Ivezi´c, ˇZ., Sesar, B., Juri´c, M., et al. 2008, The Astrophysical Journal, 684, 287 2

    2023

  19. [27]

    R., Keller, S., Frebel, A., et al

    Jacobson, H. R., Keller, S., Frebel, A., et al. 2015, The Astrophysical Journal, 807, 171 4

    2015

  20. [28]

    2022, Monthly Notices of the Royal Astronomical Society, 510, 4308 3

    Kim, B., & L´epine, S. 2022, Monthly Notices of the Royal Astronomical Society, 510, 4308 3

    2022

  21. [29]

    2011, The Messenger, 146 3

    Kuijken, K., et al. 2011, The Messenger, 146 3

    2011

  22. [30]

    2019, Astronomy & Astrophysics, 625, A2 3, 4, 13

    Kuijken, K., Heymans, C., Dvornik, A., et al. 2019, Astronomy & Astrophysics, 625, A2 3, 4, 13

    2019

  23. [31]

    2018, The Astrophysical Journal Supplement Series, 238, 16 4

    Li, H., Tan, K., & Zhao, G. 2018, The Astrophysical Journal Supplement Series, 238, 16 4

    2018

  24. [32]

    Lian, J., Bergemann, M., Pillepich, A., Zasowski, G., & Lane, R. R. 2023, Nature Astronomy, 7, 951 2

    2023

  25. [33]

    2022, Monthly Notices of the Royal Astronomical Society, 510, 433 11, 13

    Lin, J., Casagrande, L., & Asplund, M. 2022, Monthly Notices of the Royal Astronomical Society, 510, 433 11, 13

    2022

  26. [34]

    2015, Research in Astronomy and Astrophysics, 15, 1089 1

    Liu, X.-W., Zhao, G., & Hou, J.-L. 2015, Research in Astronomy and Astrophysics, 15, 1089 1

    2015

  27. [35]

    2024, The Astrophysical Journal Supplement Series, 271, 26 11, 12, 13

    Lu, X., Yuan, H., Xu, S., et al. 2024, The Astrophysical Journal Supplement Series, 271, 26 11, 12, 13

    2024

  28. [36]

    2024, ApJS, 271, 26 2 Metallicities of KiDS stars 17

    Lu, X., Yuan, H., Xu, S., et al. 2024, ApJS, 271, 26 2 Metallicities of KiDS stars 17

    2024

  29. [37]

    2015, Research in Astronomy and Astrophysics, 15, 1095 2

    Luo, A.-L., Zhao, Y .-H., Zhao, G., et al. 2015, Research in Astronomy and Astrophysics, 15, 1095 2

    2015

  30. [38]

    R., Schiavon, R

    Majewski, S. R., Schiavon, R. P., Frinchaboy, P. M., et al. 2017, The Astronomical Journal, 154, 94 1

    2017

  31. [39]

    F., Starkenburg, E., Yuan, Z., et al

    Martin, N. F., Starkenburg, E., Yuan, Z., et al. 2023, arXiv preprint arXiv:2308.01344 1, 3, 11, 12, 13 Mendes de Oliveira, C. e., Ribeiro, T., Schoenell, W., et al. 2019, Monthly Notices of the Royal Astronomical Society, 489, 241 2

    2023 arXiv

  32. [40]

    2023, Astronomy & Astrophysics, 671, A61 9

    Nepal, S., Guiglion, G., De Jong, R., et al. 2023, Astronomy & Astrophysics, 671, A61 9

    2023

  33. [41]

    L., Bovy, J., Bird, J

    Nidever, D. L., Bovy, J., Bird, J. C., et al. 2014, The Astrophysical Journal, 796, 38 2

    2014

  34. [42]

    2011, the Journal of machine Learning research, 12, 2825 6

    Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, the Journal of machine Learning research, 12, 2825 6

    2011

  35. [43]

    2013, Monthly Notices of the Royal Astronomical Society, 434, 3165 1

    Peng, X., Du, C., Wu, Z., Ma, J., & Zhou, X. 2013, Monthly Notices of the Royal Astronomical Society, 434, 3165 1

    2013

  36. [44]

    2020, Monthly Notices of the Royal Astronomical Society, 499, 1037 2

    Rojas-Arriagada, A., Zasowski, G., Schultheis, M., et al. 2020, Monthly Notices of the Royal Astronomical Society, 499, 1037 2

    2020

  37. [45]

    J., Finkbeiner, D

    Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, The Astrophysical Journal, 500, 525 4, 5, 10

    1998

  38. [46]

    2017, Monthly Notices of the Royal Astronomical Society, 471, 2587 2

    Starkenburg, E., Martin, N., Youakim, K., et al. 2017, Monthly Notices of the Royal Astronomical Society, 471, 2587 2

    2017

  39. [47]

    2006, The Astronomical Journal, 132, 1645 1

    Steinmetz, M., Zwitter, T., Siebert, A., et al. 2006, The Astronomical Journal, 132, 1645 1

    2006

  40. [48]

    2023, The Astronomical Journal, 166, 126 1, 2

    Sun, M., Chen, B., Guo, H., et al. 2023, The Astronomical Journal, 166, 126 1, 2

    2023

  41. [49]

    2015, Astronomy & Astrophysics, 575, A25 3

    Sutherland, W., Emerson, J., Dalton, G., et al. 2015, Astronomy & Astrophysics, 575, A25 3

    2015

  42. [50]

    F., Annau, N., McConnachie, A., et al

    Thomas, G. F., Annau, N., McConnachie, A., et al. 2019, arXiv preprint arXiv:1910.03076 7

    2019 arXiv

  43. [51]

    G., Prusti, T., et al

    Vallenari, A., Brown, A. G., Prusti, T., et al. 2023, Astronomy & Astrophysics, 674, A1 11, 13

    2023

  44. [52]

    2019, Monthly Notices of the Royal Astronomical Society, 482, 2189 2

    Wang, C., Liu, X., Xiang, M., et al. 2019, Monthly Notices of the Royal Astronomical Society, 482, 2189 2

    2019

  45. [53]

    D., Placco, V

    Whitten, D. D., Placco, V . M., Beers, T. C., et al. 2021, The Astrophysical Journal, 912, 147 1, 3

    2021

  46. [54]

    2017, Monthly Notices of the Royal Astronomical Society, 468, 3368 2

    Wojno, J., Kordopatis, G., Piffl, T., et al. 2017, Monthly Notices of the Royal Astronomical Society, 468, 3368 2

    2017

  47. [55]

    A., Luvaul, L

    Wolf, C., Onken, C. A., Luvaul, L. C., et al. 2018, Publications of the Astronomical Society of Australia, 35, e010 2

    2018

  48. [56]

    2014, Proceedings of the International Astronomical Union, 10, 340 4

    Wu, Y ., Du, B., Luo, A., Zhao, Y ., & Yuan, H. 2014, Proceedings of the International Astronomical Union, 10, 340 4

    2014

  49. [57]

    2024, ApJ, 968, L24 2

    Xiao, K., Huang, B., Huang, Y ., et al. 2024, ApJ, 968, L24 2

    2024

  50. [58]

    2022, ApJS, 258, 44 2

    Xu, S., Yuan, H., Niu, Z., et al. 2022, ApJS, 258, 44 2

    2022

  51. [59]

    2019, The Astrophysical Journal, 880, 36 1

    Yan, Y ., Du, C., Liu, S., et al. 2019, The Astrophysical Journal, 880, 36 1

    2019

  52. [60]

    J., et al

    Yanny, B., Rockosi, C., Newberg, H. J., et al. 2009, The Astronomical Journal, 137, 4377 1

    2009

  53. [61]

    F., et al

    Youakim, K., Starkenburg, E., Martin, N. F., et al. 2020, Monthly Notices of the Royal Astronomical Society, 492, 4986 2

    2020

  54. [62]

    2023, arXiv preprint arXiv:2311.09294 2 18 B.-K

    Zhang, H., Ardern-Arentsen, A., & Belokurov, V . 2023, arXiv preprint arXiv:2311.09294 2 18 B.-K. Sun et al

    2023 arXiv

  55. [63]

    2024, arXiv preprint arXiv:2402.06242 11, 12, 13

    Zhang, M., Xiang, M., Ting, Y .-S., et al. 2024, arXiv preprint arXiv:2402.06242 11, 12, 13

    2024 arXiv

  56. [64]

    2012, Research in Astronomy and Astrophysics, 12, 723 1

    Zhao, G., Zhao, Y .-H., Chu, Y .-Q., Jing, Y .-P., & Deng, L.-C. 2012, Research in Astronomy and Astrophysics, 12, 723 1

    2012

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.