Pith. sign in

REVIEW 5 major objections 6 minor 1 cited by

Mapping H$\alpha$-Excess Candidate Point Sources in the Southern Hemisphere Using S-PLUS Data

T0 review · 5 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read S-PLUS data yield 6,956 candidate H-alpha-excess point sources in the southern sky.

desk verdict A useful southern-sky candidate catalogue, with a load-bearing GDS locus assumption and circular ML validation that a revision should fix. read the letter →

arxiv 2501.16530 v1 pith:5W7IYVLY submitted 2025-01-27 astro-ph.GA astro-ph.SR

L. A. Gutiérrez-Soto , R. Lopes de Oliveira , S. Akras , D. R. Gonçalves , L. F. Lomelí-Núñez , C. Mendes de Oliveira , E. Telles , A. Alvarez-Candal
show 13 more authors
M. Borges Fernandes S. Daflon C. E. Ferreira Lopes M. Grossi D. Hazarika P. K. Humire C. Lima-Dias A. R. Lopes J. L. Nilo Castellón S. Panda A. Kanaan T. Ribeiro W. Schoenell
This is my paper · ORCID
classification astro-ph.GAastro-ph.SR
keywords H-alphaexcessS-PLUSsurveyJ0660narrowbandfiltercolour-colourselectionUMAPHDBSCANRRLyraeQSOandAGNclassification
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

The paper claims that the S-PLUS survey's narrowband J0660 filter, together with the broadband r and i filters, can reveal point sources whose light is modified by hydrogen-alpha emission or by other emission lines shifted into that filter. Applying a per-field stellar-locus colour excess method to the survey's fourth data release, it builds a catalogue of 6,956 candidates: 3,637 from the high-latitude Main Survey and 3,319 from the Galactic Disk Survey. Feeding all 12 S-PLUS colours to unsupervised clustering separates Galactic from extragalactic sources and isolates RR Lyrae stars, whose apparent excess arises from H-alpha absorption combined with the survey's sequential filter exposures. Adding mid-infrared colours sharpens the separation so cataclysmic variables can be told apart from quasars at redshift near 1.35, and a feature-ranking model distills the discrimination into simple colour-colour diagrams with tentative cuts. If correct, the work gives the southern sky a homogeneous H-alpha-excess candidate list and cheap pre-classification criteria for follow-up spectroscopy.

What carries the argument

The central machinery is the (r−J0660) versus (r−i) colour-colour excess criterion: a linear fit to the stellar locus per field and magnitude bin, refined by iterative sigma clipping, defines the expected colour of a source without H-alpha; the selection condition (r−J0660)obs − (r−J0660)fit ≥ 5·σest flags candidates, where σest combines the fit scatter with the photometric errors of both colours. For the Galactic Disk Survey the iterative fit is moved upward toward the unreddened main sequence to handle differential reddening. Classification uses UMAP dimensionality reduction on 66 S-PLUS colour indices followed by HDBSCAN clustering, with a second pass adding two infrared bands to make 77 colours; a Random Forest trained on the cluster labels ranks the colours and provides the basis for simplified colour-colour cuts.

What would settle it

Randomly select a few hundred candidates from each survey and take optical spectra, counting the fraction with an emission or absorption feature inside J0660; if that fraction falls far below the paper's quoted roughly 60% emission plus 30% absorption among spectral matches, the 5-sigma locus criterion is not selecting what it claims. Re-running the fit with the threshold varied from 3 to 7 sigma should also change the candidate counts smoothly, not in a cliff that would signal an unstable fitted locus.

Watch

Extended reading notes

Core claim

The central claim is that H-alpha-excess point sources over the southern sky can be identified from S-PLUS DR4 photometry and grouped into astrophysical classes using colours alone. In the (r−J0660) versus (r−i) plane, non-emitting stars define a locus; the authors fit a line to that locus in each field and magnitude bin with iterative sigma clipping, then flag any source lying at least five estimated standard deviations above the fit. In crowded Galactic-disk fields with two loci produced by differential reddening, the fit is iterated upward to the presumed unreddened main sequence. Cross-matches show the selected objects are emission-line stars, young stellar objects, binaries, cataclysmic variables, planetary nebulae, RR Lyrae stars, quasars, and active galactic nuclei; for some quasars the J0660 excess is actually Mg II, C III], C IV, [O III], or H-beta redshifted into the filter. Clustering on 66 S-PLUS colours separates a quasar/AGN-dominated group from a group of RR Lyrae stars and binaries, and adding infrared colours resolves the cataclysmic-variable/quasar confusion. The authors deliberately call the sources 'H-alpha-excess' rather than 'H-alpha emitters,' since absorption lines and variability can produce the same photometric signature.

Load-bearing premise

The candidate list rests on the fitted lines in the (r−J0660) versus (r−i) plane tracing the colours of genuinely non-emitting stars, and, in the Galactic Disk Survey, on the upper locus being the unreddened main sequence; if reddening or calibration shifts those fits, the 5-sigma criterion adds contaminants or drops real emitters.

Editorial extensions

If this is right

  • The 6,956-entry catalogue gives southern-sky surveys a compact point-source list for spectroscopic follow-up of H-alpha-excess candidates.
  • Any S-PLUS field with clean photometry can be processed with the same per-field locus fit, so the method extends to future S-PLUS data releases.
  • RR Lyrae stars and eclipsing binaries will keep appearing in sequential narrowband H-alpha selections because phase sampling and H-alpha absorption mimic excess; the paper shows how to recognise that population.
  • Quasars and AGN at particular redshifts masquerade as H-alpha emitters when Mg II, C III], C IV, [O III], or H-beta lands in J0660, so infrared colours are needed to separate extragalactic from Galactic candidates.
  • The simplified colour-colour diagrams distilled from the clustering allow pre-classification with a few filters, reducing the need for complete 12-band photometry.

Reading between the lines

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

  • The same per-field locus-fitting procedure should transfer to the northern-sky twin survey with the same filter set; a direct test is cross-matching the colour criteria in overlap regions.
  • Because the 5-sigma threshold is fixed and estimated per field, catalogue purity and completeness are likely patchy; a continuous probabilistic excess score would give each candidate an individual confidence.
  • The clean separation of RR Lyrae stars suggests the 12-filter sequence itself encodes short-period variability information, which could be mined for variability beyond the H-alpha-excess sample.
  • The remaining optical-only confusion between cataclysmic variables and quasars at z ~ 1.35 marks a boundary for photometric classification that infrared colours only partly resolve; adding other line-sensitive bands could push it further.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper presents a search for Hα-excess point sources in S-PLUS DR4 using the (r−J0660) versus (r−i) colour-colour method of Witham et al. (2008). The authors first apply iterative linear fits to the stellar locus in each field and magnitude bin for the Main Survey (MS) and the Galactic Disk Survey (GDS), selecting sources that lie at least 5σ above the fit. They report 6,956 candidates, 3,637 from the MS and 3,319 from the GDS, and cross-match them with SIMBAD, SDSS, LAMOST, and VPHAS+. They then use UMAP plus HDBSCAN to cluster the MS candidates, first with 66 S-PLUS colours and then with 77 colours including WISE W1/W2, and train a Random Forest on the cluster labels to identify important colours. The paper concludes that the S-PLUS 12-filter system can separate RR Lyrae stars and eclipsing binaries from genuine Hα emitters, that optical plus infrared data improve the separation of Galactic from extragalactic sources, and that the proposed colour-colour diagrams provide tentative pre-classification criteria.

Significance. If the candidate catalogue is sound, this work would provide a new southern-sky list of Hα-excess point sources and demonstrate the utility of S-PLUS narrow-band J0660 photometry for finding emission-line objects and variables. The paper has several genuine strengths: the selection procedure is described in enough detail to be followed; the SDSS and LAMOST spectra in Figs. 7–10 convincingly show that at least some selected objects are genuine emitters, CVs, and QSOs with redshifted lines; and the authors are explicit about contamination by RR Lyrae and eclipsing binaries and about caution for r<13.5 mag. However, the central catalogue is not actually included or linked, the GDS selection depends on an unverified choice of the 'upper locus' as the unreddened main sequence, and the Random Forest validation is circular with respect to physical classification because it is trained on HDBSCAN labels generated from the same photometric features. These issues currently prevent the reader from independently checking the headline numbers or the physical interpretation of the clusters.

major comments (5)
  1. [3.3 (Eq. 1, Fig. 5)] The GDS upper-locus fit is the load-bearing step for the 3,319 GDS candidates. The text states that the upper locus 'generally corresponds to the unreddened main sequence' but never demonstrates this by overlaying the Pickles main-sequence and giant tracks shown in Figs. 2–3 on a GDS field. Because the iterative procedure re-fits points above the initial line, genuine Hα emitters can drag the fitted line upward, and if the upper locus is in fact a reddened or giant-dominated population, Eq. (1) thresholds shift for every GDS source. The procedure is also not fully specified: the clipping threshold is described only as 'several σ' and the iteration count as '2 to 3'. Please overlay the Pickles tracks on representative GDS fields, quantify the offset between the initial and final fits per field and magnitude bin, specify the clipping algorithm precisely, and report how the 3,319-candidate count changes if the lower (reddened) locus is used as the reference.
  2. [4 and 6] The central catalogue is not provided. The headline result of the paper is a list of 6,956 Hα-excess candidates, but the manuscript contains no machine-readable table, no VizieR/CDS link, and no appendix table with candidate positions, photometry, errors, and flags. Without the catalogue, readers cannot verify the reported counts, the SIMBAD cross-match statistics, or the clustering metadata. Please add a full catalogue, ideally at CDS/VizieR, and describe its columns, duplicate handling, and provenance.
  3. [5.3 (Fig. 18, Fig. 19, Table 3)] The Random Forest validation is circular with respect to physical classification. The model is trained on HDBSCAN labels generated from the same 66–77 photometric colours used as input features, so the reported F1 macro average of 0.95 and the feature importances quantify how well the Random Forest reproduces the cluster labels, not whether the clusters correspond to physically distinct source classes. This circularity also affects the interpretation of Groups 0–4 in Table 3 and the tentative colour cuts in Fig. 19. Please either reframe Section 5.3 as an analysis of cluster-label reproducibility or validate the clusters against external spectroscopic classifications, reporting per-class precision and recall for confirmed CVs, QSOs, RR Lyrae stars, and eclipsing binaries.
  4. [4.3] The spectroscopic confirmation rates are not defined on a well-posed sample. The paper states that 'approximately 60%' of the 212 SDSS/LAMOST matches show emission lines and 'around 30%' show Hα absorption, but the denominator, the selection function, and any spectral signal-to-noise or line-strength thresholds are not given. The SDSS and LAMOST overlaps are not a random subsample of the candidate list, so these percentages cannot be interpreted as purity estimates for the full catalogue. Please provide a per-survey table of spectral classifications (emission, absorption, no line, unknown) and, if possible, report confirmation rates for a randomly selected or completeness-weighted subsample.
  5. [4.4 (Fig. 11)] The VPHAS+ comparison does not validate the candidate selection. The 793 matched objects are used to compare r−i and r−Hα colour differences between S-PLUS and VPHAS+, yielding medians of −0.21 and +0.02 with MADs of 0.07 and 0.27, but this does not test whether S-PLUS Hα-excess candidates coincide with VPHAS+ Hα-excess sources. A meaningful external check would cross-match the S-PLUS GDS candidate list with VPHAS+ photometry and compare the Hα-excess flags, reporting the fraction of S-PLUS candidates that are also VPHAS+ emitters and vice versa.
minor comments (6)
  1. [Header] The manuscript header contains placeholder dates ('Received September 15, 1996; accepted March 16, 1997') that should be corrected before submission.
  2. [Section 4 and Fig. 6] The filter is repeatedly written as 'J0600' in Section 4 and in the caption of Fig. 6; it should be 'J0660' consistently throughout the paper.
  3. [Fig. 19] The colour-colour diagrams in Fig. 19 are difficult to use because the pairplot panels lack clear axis labels and the tentative colour cuts are not legible in the printed figure; please replot with labelled axes and a legend for the cuts.
  4. [Section 5.2.1] The sentence 'We ended up with the optimal value of 2 and 50, respectively' does not specify which HDBSCAN parameter is 2 and which is 50; the order for min_samples and min_cluster_size should be stated explicitly.
  5. [Section 5.2.2] The text 'typically exhibit amplitudes higher than 0.3 to 2 magnitudes for RR Lyrae stars Chandra X-ray Observatory' contains what appears to be a citation placeholder; please replace it with proper references for RR Lyrae periods and amplitudes.
  6. [Section 2] There are several grammar and typographical slips, such as 'we primary goal of identify objects' in Section 2 and 'dmax,pro j' in Section 4; these should be corrected in a copy-editing pass.

Circularity Check

1 steps flagged · score 6.0 of 10

The Hα-excess catalogue is an independent empirical selection, but the machine-learning feature-importance and tentative colour-cut results are circular because the Random Forest is trained on HDBSCAN labels generated from the same photometric colours it is then used to 'discover'.

  1. fitted input called prediction [Section 5.3, 'Extracting Main Features: Colour Analysis', and Figure 19]
    "We used the classifications made by combining UMAP and HDBSCAN to create Random Forest (Breiman 2001) models and identified the most important features, specifically the colours that contribute to the separation or classification of the classes of objects. ... using 66 S-PLUS colours plus 11 additional colours generated with the W1 and W2 filters as input parameters, and labels generated by HDBSCAN. ... Tentative colour cuts are presented in the Fig. 19 to differentiate the various classes of Hα sources."

    The HDBSCAN cluster labels are produced by UMAP embeddings of the same 66 (or 77) photometric colours that are later used as input features to the Random Forest. The labels therefore contain no information beyond the input features, so the reported F1 Macro Average of 0.95 measures only how well a flexible model can re-learn partitions created from the same data, not an independent predictive validation. The feature importances and the tentative colour cuts in Figure 19 are drawn to enclose these same clusters, so presenting them as 'key features' or 'tentative criteria' for classifying Hα-excess sources is a description of the clustering input rather than an independent discovery about the classes.

full rationale

The primary catalogue is not circular: Eq. (1) defines J0660 excess relative to a per-field, per-magnitude-bin stellar-locus fit, and the resulting 6,956 candidates are an empirical selection whose interpretation is cross-checked against external data (SIMBAD, SDSS, LAMOST, VPHAS+). The GDS upper-locus choice is an assumption with possible systematics, but it is not defined in terms of the output and therefore is a correctness risk rather than a circularity. The circular step is confined to the machine-learning feature analysis: the Random Forest is trained on HDBSCAN cluster labels that were generated from the same photometric colours used as RF inputs, so its high F1 score, the feature importances, and the tentative colour cuts only recover the previously imposed cluster structure. This gives a partial circularity score of 6: the central candidate catalogue has independent content, but one of the paper's advertised products, the simplified colour classification, reduces by construction.

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

The catalogue depends on a string of hand-chosen thresholds (C=5, magnitude bins, quality cuts, ML hyperparameters) and on the assumption that the fitted stellar locus is a clean separator between normal stars and J0660-excess sources. No new physical entities are introduced; the clustering labels are internal to the data.

free parameters (5)
  • Excess threshold C = 5
    Eq. (1) fixes C=5 in (r-J0660)obs - (r-J0660)fit >= C*sigma_est; the value is adopted without a derivation and directly controls the size and purity of the candidate catalogue.
  • Photometric and morphological selection thresholds = r in (13,19.5]; J0660<19.4; i<19.2; errors<0.2 mag; S/N>10; CLASS_STAR=1; ellipticity<0.2; FLUX_RADIUS_50<3
    Section 3.1 lists hand-chosen cuts that define which sources enter the colour-colour analysis; different cuts would change the candidate list.
  • Magnitude bin boundaries = 13-16, 16-17.5, 17.5-18.5, 18.5-19.5
    Section 3.3 divides the sample into four r-band bins to keep photometric scatter comparable; the edges are arbitrary and affect both the fits and the number of selected candidates.
  • UMAP hyperparameters = S-PLUS-only: n_neighbors=30, n_components=2, min_dist=0.1; with WISE: n_neighbors=50, n_components=2
    Section 5.2.1 and 5.2.2 choose these by maximizing silhouette score and Davies-Bouldin index on the same data, so the clusters are partly shaped by these choices.
  • HDBSCAN hyperparameters = S-PLUS-only: min_cluster_size=2, min_samples=50; with WISE: min_cluster_size=50, min_samples=5
    Section 5.2 states these were chosen after experimentation; the resulting group memberships depend on them.
assumptions (5)
  • domain assumption The stellar locus in (r-J0660) versus (r-i) is approximately linear in each field and magnitude bin, and the upper locus used for the GDS is the unreddened main sequence.
    Section 3.3 uses linear regression plus iterative sigma clipping to define the locus; if reddening or multiple populations break linearity, Eq. (1) misidentifies excess sources.
  • domain assumption Photometric deviations from the fitted locus are dominated by intrinsic J0660 line excess or variability, not by residual calibration errors or reddening.
    The selection labels all C-sigma outliers as H-alpha-excess candidates; the paper itself acknowledges RR Lyrae and binary variability create false excesses, so this assumption is only partially satisfied.
  • domain assumption SIMBAD classifications and the SDSS/LAMOST spectra used for validation correctly identify the nature of matched objects.
    Sections 4.1 and 4.3 use those classifications to label clusters and to quote the 60% emission and 30% absorption rates; classification errors propagate into purity estimates.
  • domain assumption S-PLUS DR4 flux calibration is uniform across fields except for the five GDS fields excluded by the authors.
    Section 3.2 removes five fields with apparent zero-point offsets; the rest are assumed calibrated to the stated 0.01-0.02 mag accuracy.
  • standard math The UMAP and HDBSCAN algorithms behave as described by their cited references when applied to photometric colours.
    The clustering results inherit whatever approximations or instabilities are present in the UMAP manifold construction and HDBSCAN density estimates.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mapping H$\alpha$-Excess Candidate Point Sources in the Southern Hemisphere Using S-PLUS Data." pith.science (2026). https://pith.science/paper/5W7IYVLY

@misc{pith2026250116530,
  author       = {Pith},
  title        = {Pith review of: Mapping H$\alpha$-Excess Candidate Point Sources in the Southern Hemisphere Using S-PLUS Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5W7IYVLY}},
  note         = {Machine review of arXiv:2501.16530}
}
abstract

Context. We use the Southern Photometric Local Universe Survey (S-PLUS) Fourth Data Release (DR4) to identify and classify H$\alpha$-excess point sources in the Southern Sky, combining photometric data from 12 S-PLUS filters with machine learning to improve classification of H$\alpha$-related phenomena. Aims. Our goal is to classify H$\alpha$-excess point sources by distinguishing Galactic and extragalactic objects, particularly those with redshifted emission lines, and identifying variability phenomena like RR Lyrae stars. Methods. We selected H$\alpha$-excess candidates using the ($r - J0660$) vs. ($r - i$) colour-colour diagram from the S-PLUS main survey (MS) and Galactic Disk Survey (GDS). UMAP for dimensionality reduction and HDBSCAN clustering were used to separate source types. Infrared data was incorporated, and a Random Forest model was trained on clustering results to identify key colour features. New colour-colour diagrams from S-PLUS MS and infrared data offer a preliminary classification. Results. Combining multiwavelength data with machine learning significantly improved H$\alpha$-excess source classification. We identified 6956 sources with excess in the $J0660$ filter. Cross-matching with SIMBAD explored object types, including emission-line stars, young stellar objects, nebulae, stellar binaries, cataclysmic variables, QSOs, AGNs, and galaxies. Using S-PLUS colours and machine learning, we separated RR Lyrae stars from other sources. The separation of Galactic and extragalactic sources was clearer, but distinguishing cataclysmic variables from QSOs at certain redshifts remained challenging. Infrared data refined the classification, and the Random Forest model highlighted key colour features for future follow-up spectroscopy.

Figures

Figures reproduced from arXiv: 2501.16530 by the authors.

Figure 1
Figure 1. Transmission curves of the S-PLUS filter set. The narrowband filter J0660 includes the Hα emission line. Over-plotted is spectra of different classes of emission line objects. From top to bottom: a PN, a symbiotic star, an extragalactic H ii region, a blue compact/H ii galaxy, a YSO, a CV star, a B[e] star, a star-forming galaxy and a QSO at a redshift of ∼3.31. – Errors less than 0.2 mag in the r, J0660, and i filt… view at source ↗
Figure 2
Figure 2. The (r − J0660) versus (r − i) colour-colour plots used to select objects with Hα excess. These plots display data for all stars from the S-PLUS DR4 MS, representing the PStotal photometry in these colours. The data are divided into four magnitude bins: (a) 13 < r ≤ 16, (b) 16 < r ≤ 17.5, (c) 17.5 < r ≤ 18.5, and (d) 18.5 < r ≤ 19.5. Objects with Hα excess are expected to be located towards the top of these diagrams… view at source ↗
Figure 3
Figure 3. Same as [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (15 more)
Figure 4
Figure 4. Figure 4: Illustration of the selection criteria used to identify strong emission-line objects via colour-colour plots. The data shown here are from the S-PLUS field STRIPE82-0142, split into four magnitude bins, as displayed in the four panels. The thin red continuous lines sho…
Figure 5
Figure 5. Figure 5: Colour-colour diagram of the stars locate at field SPLUS-d288 like those found in [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: The colour-colour diagram shows the distribution of Hα￾feature sources in the (r− J0660) versus (r−i) colour-colour space. The data is divided into two populations: GDS and MS representing distinct galactic components. The GDS population, depicted by filled circles in …
Figure 7
Figure 7. Figure 7: presents the SDSS (upper) and LAMOST (lower) spectra, along with the corresponding S-PLUS photometry (col￾oured symbols) for two known cataclysmic variables (CVs) and one eclipsing binary, respectively. The excess in the J0660 fil￾ter is evidently produced by the Hα li…
Figure 8
Figure 8. Figure 8: SDSS spectrum and S-PLUS photometry of the RR Lyrae star SDSS J010045.13-010212.2, showing an Hα absorption line. 3000 4000 5000 6000 7000 8000 9000 Wavelength (Å) 0.0 0.2 0.4 0.6 F(1 0 1 5 e r g s 1 c m 2 Å 1 ) Mg II 2799 QSO z = 1.359 3000 4000 5000 6000 7000 8000 90…
Figure 10
Figure 10. Figure 10: As [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 9
Figure 9. Figure 9: S-PLUS photometry and SDSS spectra of three QSOs with redshifts of 1.359, 2.454 and 3.280 (top to bottom) selected as Hα ex￾cess sources. At these redshifts, the emission lines Mg II λ2799, C III] λ1909 and C IV λ1551 are detected in the J0660 filter. The SDSS IDs of t…
Figure 11
Figure 11. Figure 11: Histograms illustrating the discrepancies in the photometric colours r − i and r - Hα between the S-PLUS and VPHAS+ surveys. The left panel depicts the differences in the r − i colour, while the right panel shows the differences in the r - Hα colour. Both histograms r…
Figure 12
Figure 12. Figure 12: Upper panel: Distribution of r-band magnitudes for Hα ex￾cess sources (blue curve) compared to all the stars (salmon curve) in the MS. Lower panel: Distribution of r-band magnitudes for Hα excess sources (blue curve) in the GDS compared to all stars (salmon curve). 5.…
Figure 14
Figure 14. Figure 14: Distribution of the objects in galactic longitude for Hα excess sources (blue bars) and all stars (salmon bars) for the MS (left panel) and the GDS (right panel). representation that retains the intricate nonlinear relationships present in the original high-dimensiona…
Figure 15
Figure 15. Figure 15: Silhouette Score (left panel) and Davies-Bouldin Index (right panel) as functions of the number of neighbours (n_neighbors) for different values of UMAP components (n_components). Higher Silhouette Score values and lower Davies-Bouldin Index values indicate better clu…
Figure 16
Figure 16. Figure 16: UMAP dimension reduction applied to the MS from S-PLUS data. The left panel shows the UMAP result using only the S-PLUS colours as input parameters, with the colour bar indicating the r magnitude. The right panel displays the result after applying HDBSCAN clustering, …
Figure 17
Figure 17. Figure 17: Similar to [PITH_FULL_IMAGE:figures/full_fig_p016_17.png]
Figure 18
Figure 18. Figure 18: Top 20 feature importances identified by the Random Forest model, showing the colours that contributed most significantly to the clustering of Hα-excess objects using UMAP + HDBSCAN. The im￾portance scores indicate the relative impact of each colour on the clas￾sifica…
Figure 19
Figure 19. Figure 19: Examples of colour-colour diagrams using the top 20 features identified by the Random Forest model. These diagrams show the separation of different classes of objects from the Hα-excess sources list. The selected diagrams illustrate effective clustering achieved throu…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Symbiotic stars in the era of modern ground- and space-based surveys

    astro-ph.SR 2025-04 accept novelty 2.0 of 10

    A review of how large-scale photometric and spectroscopic surveys have transformed the discovery, classification, and variability studies of symbiotic stars.

Reference graph

Works this paper leans on

96 extracted references · 48 canonical work pages · cited by 1 Pith paper

  1. [1]

    2020, , 492, L40

    Abril , J., Schmidtobreick , L., Ederoclite , A., & L \'o pez-Sanjuan , C. 2020, , 492, L40

    2020

  2. [2]

    A., Almeida , A., et al

    Ahumada , R., Prieto , C. A., Almeida , A., et al. 2020, , 249, 3

    2020

  3. [3]

    2023, , 519, 6044

    Akras , S. 2023, , 519, 6044

    2023

  4. [4]

    R., Alvarez-Candal , A., & Pereira , C

    Akras , S., Gon c alves , D. R., Alvarez-Candal , A., & Pereira , C. B. 2021, , 502, 2513

    2021

  5. [5]

    Akras , S., Guzman-Ramirez , L., & Gon c alves , D. R. 2019 a , , 488, 3238

    2019

  6. [6]

    L., & Ramos-Larios , G

    Akras , S., Guzman-Ramirez , L., Leal-Ferreira , M. L., & Ramos-Larios , G. 2019 b , , 240, 21

    2019

  7. [7]

    L., Guzman-Ramirez , L., & Ramos-Larios , G

    Akras , S., Leal-Ferreira , M. L., Guzman-Ramirez , L., & Ramos-Larios , G. 2019 c , , 483, 5077

    2019

  8. [8]

    R., et al

    Almeida-Fernandes , F., SamPedro , L., Herpich , F. R., et al. 2022, , 511, 4590

    2022

Show all 96 references
  1. [9]

    J., Drew , J

    Barentsen , G., Farnhill , H. J., Drew , J. E., et al. 2014, , 444, 3230

    2014

  2. [10]

    S., Drew , J

    Barentsen , G., Vink , J. S., Drew , J. E., et al. 2011, , 415, 103

    2011

  3. [11]

    2018, Nature biotechnology

    Becht, E., McInnes, L., Healy, J., et al. 2018, Nature biotechnology

    2018

  4. [12]

    2014, arXiv e-prints, arXiv:1403.5237

    Benitez , N., Dupke , R., Moles , M., et al. 2014, arXiv e-prints, arXiv:1403.5237

    2014 arXiv

  5. [13]

    2011, in Astronomical Society of the Pacific Conference Series, Vol

    Bertin , E. 2011, in Astronomical Society of the Pacific Conference Series, Vol. 442, Astronomical Data Analysis Software and Systems XX, ed. I. N. Evans , A. Accomazzi , D. J. Mink , & A. H. Rots , 435

    2011

  6. [14]

    & Arnouts , S

    Bertin , E. & Arnouts , S. 1996, , 117, 393

    1996

  7. [15]

    Blair , W. P. & Long , K. S. 2004, , 155, 101

    2004

  8. [16]

    R., Cortesi , A., Lucatelli , G., et al

    Bom , C. R., Cortesi , A., Lucatelli , G., et al. 2021, , 507, 1937

    2021

  9. [17]

    2021, , 653, A31

    Bonoli , S., Mar \' n-Franch , A., Varela , J., et al. 2021, , 653, A31

    2021

  10. [18]

    2001, Machine Learning, 45, 5

    Breiman, L. 2001, Machine Learning, 45, 5

    2001

  11. [19]

    Campello, R. J. G. B., Moulavi, D., & Sander, J. 2013, in Advances in Knowledge Discovery and Data Mining, ed. J. Pei, V. S. Tseng, L. Cao, H. Motoda, & G. Xu (Berlin, Heidelberg: Springer Berlin Heidelberg), 160--172

    2013

  12. [20]

    J., Moles , M., Crist \'o bal-Hornillos , D., et al

    Cenarro , A. J., Moles , M., Crist \'o bal-Hornillos , D., et al. 2019, , 622, A176

    2019

  13. [21]

    Coelho , P. R. T. 2014, , 440, 1027

    2014

  14. [22]

    O., Kasliwal , M

    Cook , D. O., Kasliwal , M. M., Van Sistine , A., et al. 2019, , 880, 7

    2019

  15. [23]

    Corradi , R. L. M. & Giammanco , C. 2010, , 520, A99

    2010

  16. [24]

    Corradi , R. L. M., Rodr \' guez-Flores , E. R., Mampaso , A., et al. 2008, , 480, 409

    2008

  17. [25]

    Corradi , R. L. M., Sabin , L., Munari , U., et al. 2011, , 529, A56

    2011

  18. [26]

    M., Wright , E

    Cutri , R. M., Wright , E. L., Conrow , T., et al. 2013, Explanatory Supplement to the AllWISE Data Release Products , Explanatory Supplement to the AllWISE Data Release Products, by R. M. Cutri et al

    2013

  19. [27]

    Davies, D. L. & Bouldin, D. W. 1979, IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI-1, 224

    1979

  20. [28]

    D., Elliott , K

    Davies , R. D., Elliott , K. H., & Meaburn , J. 1976, , 81, 89

    1976

  21. [29]

    E., Gonzalez-Solares , E., Greimel , R., et al

    Drew , J. E., Gonzalez-Solares , E., Greimel , R., et al. 2014, , 440, 2036

    2014

  22. [30]

    E., Greimel , R., Irwin , M

    Drew , J. E., Greimel , R., Irwin , M. J., et al. 2005, , 362, 753

    2005

  23. [31]

    E., Greimel , R., Irwin , M

    Drew , J. E., Greimel , R., Irwin , M. J., & Sale , S. E. 2008, , 386, 1761

    2008

  24. [32]

    1996, in Proc

    Ester, M., Kriegel, H.-P., Sander, J., & Xu, X. 1996, in Proc. of 2nd International Conference on Knowledge Discovery and Data Mining (KDD-96), 226--231

    1996

  25. [33]

    E., et al

    Fratta , M., Scaringi , S., Drew , J. E., et al. 2021, , 505, 1135

    2021

  26. [34]

    Frew , D. J. 2008, PhD thesis, Department of Physics, Macquarie University, NSW 2109, Australia

    2008

  27. [35]

    E., et al

    Fukugita , M., Ichikawa , T., Gunn , J. E., et al. 1996, , 111, 1748

    1996

  28. [36]

    A., Lomel \' -N \'u \ n ez , L., \'A lamo-Mart \' nez , K., et al

    Gonz \'a lez-L \'o pezlira , R. A., Lomel \' -N \'u \ n ez , L., \'A lamo-Mart \' nez , K., et al. 2017, , 835, 184

    2017

  29. [37]

    A., Lomel \' -N \'u \ n ez , L., Ordenes-Brice \ n o , Y., et al

    Gonz \'a lez-L \'o pezlira , R. A., Lomel \' -N \'u \ n ez , L., Ordenes-Brice \ n o , Y., et al. 2022, , 941, 53

    2022

  30. [38]

    A., Payne , S

    Greer , P. A., Payne , S. G., Norton , A. J., et al. 2017, , 607, A11

    2017

  31. [39]

    A., Gon c alves , D

    Guti \'e rrez-Soto , L. A., Gon c alves , D. R., Akras , S., et al. 2020, , 633, A123

    2020

  32. [40]

    A., Mari , M

    Guti \'e rrez-Soto , L. A., Mari , M. B., Weidmann , W. A., & Faifer , F. R. 2024, , 109, 102207

    2024

  33. [41]

    R., Almeida-Fernandes , F., Oliveira Schwarz , G

    Herpich , F. R., Almeida-Fernandes , F., Oliveira Schwarz , G. B., et al. 2024, , 689, A249

    2024

  34. [42]

    H., Kronberger , M., Patchick , D., et al

    Jacoby , G. H., Kronberger , M., Patchick , D., et al. 2010, , 27, 156

    2010

  35. [43]

    & Omar , A

    Jaiswal , S. & Omar , A. 2016, , 462, 92

    2016

  36. [44]

    M., Vink , J

    Kalari , V. M., Vink , J. S., Drew , J. E., et al. 2015, , 453, 1026

    2015

  37. [45]

    Kron , R. G. 1980, , 43, 305

    1980

  38. [46]

    1982, IEEE Transactions on Information Theory, 28, 129

    Lloyd, S. 1982, IEEE Transactions on Information Theory, 28, 129

    1982

  39. [47]

    D., Rodr \' guez-Merino , L

    Lomel \' -N \'u \ n ez , L., Mayya , Y. D., Rodr \' guez-Merino , L. H., Ovando , P. A., & Rosa-Gonz \'a lez , D. 2022, , 509, 180

    2022

  40. [48]

    2012, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol

    Mar \' n-Franch , A., Chueca , S., Moles , M., et al. 2012, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 8450, Modern Technologies in Space- and Ground-based Telescopes and Instrumentation II, ed. R. Navarro , C. R. Cunningham , & E. Pri...

    2012

  41. [49]

    2017, The Journal of Open Source Software, 2

    McInnes, L., Healy, J., & Astels, S. 2017, The Journal of Open Source Software, 2

    2017

  42. [50]

    2020, UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction

    McInnes, L., Healy, J., & Melville, J. 2020, UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction

    2020

  43. [51]

    2019, , 489, 241

    Mendes de Oliveira , C., Ribeiro , T., Schoenell , W., et al. 2019, , 489, 241

    2019

  44. [52]

    2019, Eruptive Stars Information Letter, 41, 78

    Merc , J., G \`a lis , R., & Wolf , M. 2019, Eruptive Stars Information Letter, 41, 78

    2019

  45. [53]

    2022, , 510, 1404

    Merc , J., G \'a lis , R., Wolf , M., et al. 2022, , 510, 1404

    2022

  46. [54]

    2021, , 506, 4151

    Merc , J., G \'a lis , R., Wolf , M., et al. 2021, , 506, 4151

    2021

  47. [55]

    2020, , 644, A49

    Merc , J., Miko ajewska , J., Gromadzki , M., et al. 2020, , 644, A49

    2020

  48. [56]

    Miko ajewska , J., Caldwell , N., & Shara , M. M. 2014, , 444, 586

    2014

  49. [57]

    M., Caldwell , N., I kiewicz , K., & Zurek , D

    Miko ajewska , J., Shara , M. M., Caldwell , N., I kiewicz , K., & Zurek , D. 2017, , 465, 1699

    2017

  50. [58]

    Miszalski , B., Acker , A., Moffat , A. F. J., Parker , Q. A., & Udalski , A. 2009, , 496, 813

    2009

  51. [59]

    & Miko ajewska , J

    Miszalski , B. & Miko ajewska , J. 2014, , 440, 1410

    2014

  52. [60]

    E., et al

    Mongui \'o , M., Greimel , R., Drew , J. E., et al. 2020, , 638, A18

    2020

  53. [61]

    M., Frasca , A., et al

    Munari , U., Alcal \'a , J. M., Frasca , A., et al. 2022, , 661, A124

    2022

  54. [62]

    2021, , 505, 6121

    Munari , U., Traven , G., Masetti , N., et al. 2021, , 505, 6121

    2021

  55. [63]

    Nakazono , L., Mendes de Oliveira , C., Hirata , N. S. T., et al. 2021, , 507, 5847

    2021

  56. [64]

    Oke , J. B. & Gunn , J. E. 1983, , 266, 713

    1983

  57. [65]

    A., Boji c i \'c , I

    Parker , Q. A., Boji c i \'c , I. S., & Frew , D. J. 2016, in Journal of Physics Conference Series, Vol. 728, Journal of Physics Conference Series, 032008

    2016

  58. [66]

    A., Phillipps , S., Pierce , M

    Parker , Q. A., Phillipps , S., Pierce , M. J., et al. 2005, , 362, 689

    2005

  59. [67]

    2011, Journal of Machine Learning Research, 12, 2825

    Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, Journal of Machine Learning Research, 12, 2825

    2011

  60. [68]

    M., Richards , G

    Peters , C. M., Richards , G. T., Myers , A. D., et al. 2015, , 811, 95

    2015

  61. [69]

    Pickles , A. J. 1998, , 110, 863

    1998

  62. [70]

    D., Vollmann , W., & Somogyi , P

    Pollmann , E., Bennett , P. D., Vollmann , W., & Somogyi , P. 2018, Information Bulletin on Variable Stars, 6249, 1

    2018

  63. [71]

    E., Steeghs , D., et al

    Raddi , R., Drew , J. E., Steeghs , D., et al. 2015, , 446, 274

    2015

  64. [72]

    Rousseeuw, P. J. 1987, Journal of Computational and Applied Mathematics, 20, 53

    1987

  65. [73]

    A., Wareing , C., et al

    Sabin , L., Zijlstra , A. A., Wareing , C., et al. 2010, , 27, 166

    2010

  66. [74]

    J., Verbeek , K., et al

    Scaringi , S., Groot , P. J., Verbeek , K., et al. 2013, , 428, 2207

    2013

  67. [75]

    & Lapalme, G

    Sokolova, M. & Lapalme, G. 2009, Information Processing & Management, 45, 427

    2009

  68. [76]

    Taylor , M. B. 2005, in Astronomical Society of the Pacific Conference Series, Vol. 347, Astronomical Data Analysis Software and Systems XIV, ed. P. Shopbell , M. Britton , & R. Ebert , 29

    2005

  69. [77]

    Viironen , K., Mampaso , A., Corradi , R. L. M., et al. 2009, , 502, 113

    2009

  70. [78]

    S., Drew , J

    Vink , J. S., Drew , J. E., Steeghs , D., et al. 2008, , 387, 308

    2008

  71. [79]

    Waskom, M. L. 2021, Journal of Open Source Software, 6, 3021

    2021

  72. [80]

    G., Nelemans , G., et al

    Wevers , T., Jonker , P. G., Nelemans , G., et al. 2017, , 466, 163

    2017

  73. [81]

    R., Knigge , C., Aungwerojwit , A., et al

    Witham , A. R., Knigge , C., Aungwerojwit , A., et al. 2007, , 382, 1158

    2007

  74. [82]

    R., Knigge , C., Drew , J

    Witham , A. R., Knigge , C., Drew , J. E., et al. 2008, , 384, 1277

    2008

  75. [83]

    R., Knigge , C., G \"a nsicke , B

    Witham , A. R., Knigge , C., G \"a nsicke , B. T., et al. 2006, , 369, 581

    2006

  76. [84]

    L., Eisenhardt , P

    Wright , E. L., Eisenhardt , P. R. M., Mainzer , A. K., et al. 2010, , 140, 1868

    2010

  77. [85]

    L., Li , H.-N., et al

    Wu , Y., Luo , A. L., Li , H.-N., et al. 2011, Research in Astronomy and Astrophysics, 11, 924

    2011

  78. [86]

    2022, , 659, A181

    Yang , L., Yuan , H., Xiang , M., et al. 2022, , 659, A181

    2022

  79. [87]

    G., Adelman , J., Anderson , John E., J., et al

    York , D. G., Adelman , J., Anderson , John E., J., et al. 2000, , 120, 1579

    2000

  80. [88]

    1966, in Stellar Evolution, ed.\ R

    Baker, N. 1966, in Stellar Evolution, ed.\ R. F. Stein,& A. G. W. Cameron (Plenum, New York) 333

    1966

  81. [89]

    1988, A&A, 200, 58

    Balluch, M. 1988, A&A, 200, 58

    1988

  82. [90]

    Cox, J. P. 1980, Theory of Stellar Pulsation (Princeton University Press, Princeton) 165

    1980

  83. [91]

    N.,& Stewart, J

    Cox, A. N.,& Stewart, J. N. 1969, Academia Nauk, Scientific Information 15, 1

    1969

  84. [92]

    1980, Prog

    Mizuno H. 1980, Prog. Theor. Phys., 64, 544

    1980

  85. [93]

    Tscharnuter W. M. 1987, A&A, 188, 55

    1987

  86. [94]

    1992, in ASP Conf

    Terlevich, R. 1992, in ASP Conf. Ser. 31, Relationships between Active Galactic Nuclei and Starburst Galaxies, ed. A. V. Filippenko, 13

    1992

  87. [95]

    Yorke, H. W. 1980a, A&A, 86, 286

  88. [96]

    F., Tytler, D

    Zheng, W., Davidsen, A. F., Tytler, D. & Kriss, G. A. 1997, preprint

    1997

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.