Pith. sign in

REVIEW 3 major objections 5 minor 41 references

Fair for a few: Improving Fairness in Doubly Imbalanced Datasets

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper proposes a three-parameter resampling search that finds data balances making fraud classifiers fairer without sacrificing accuracy.

desk verdict The three-parameter resampling idea is clean and interpretable, but the headline fairness gains are fitted to the tuning partitions and largely vanish on the one truly held-out test, so the central claim is not yet established. read the letter →

arxiv 2506.14306 v1 pith:VRVYAOPI submitted 2025-06-17 cs.LG cs.CY

Ata Yalcin , Asli Umay Ozturk , Yigit Sever , Viktoria Pauw , Stephan Hachinger , Ismail Hakki Toroslu , Pinar Karagoz This is my paper · ORCID
classification cs.LGcs.CY
keywords algorithmicfairnessdoublyimbalanceddatasetsfrauddetectionresamplinggridsearchParetofrontdisparateimpactratioMatthewscorrelationcoefficient
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 addresses datasets that are imbalanced twice over: the fraud label is rare and one sensitive-attribute group is much smaller than the other. It argues that off-the-shelf debiasing methods fail on such doubly imbalanced data, and proposes a model-agnostic resampling solution that searches for the best balance between the two privilege groups and the two labels. The search is controlled by three interpretable parameters, and the outcome is a Pareto front showing the trade-off between fairness, measured by disparate impact ratio, and accuracy, measured by Matthews correlation coefficient. The authors demonstrate on three fraud datasets that the resulting balance brings the DI ratio close to 1 for several classifiers, including when combined with an existing debiasing method, without sacrificing and sometimes improving MCC.

What carries the argument

The load-bearing object is the three-parameter balance structure $(\alpha, \beta, \gamma)$. Given the original four cell counts, privileged-favourable, privileged-unfavourable, unprivileged-favourable, and unprivileged-unfavourable, these parameters determine the four sampling ratios through a closed-form interpolation, with $\alpha$ moving the privilege split toward 50/50, $\beta$ moving the label split toward 50/50, and $\gamma$ moving the favourable-label rates of the two groups toward equality. A two-level grid search, first with coarse steps of 0.1 and then with refinement around the top five points at steps of 0.01, evaluates the combined loss for every parameter triple and for each classifier's probability threshold. The Pareto front of the two loss components is reported so that a user can pick the operating point that best fits their fairness and accuracy requirements.

What would settle it

Take the BAF dataset, hold out a test partition that is never touched by the grid search, and re-run the full two-level search on several independent 5,000-row training samples; if the selected $(\alpha, \beta, \gamma)$ values vary widely across samples, or if the DI ratio on the untouched test set falls below 0.8 while the tuning-partition DI ratio stays near 1, then the reported fairness gain is an artifact of tuning on the evaluation sample.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that the optimal training distribution for a fair classifier on a doubly imbalanced dataset is not a fully balanced sample but a dataset-specific intermediate point that can be found by search. The proposed method defines the training sample by three parameters: $\alpha$ controls the proportion of the unprivileged group, $\beta$ controls the proportion of the unfavourable label, and $\gamma$ controls the ratio of favourable-label rates between privileged and unprivileged groups. A two-level grid search evaluates the combined loss $c_1 \cdot |1 - \mathrm{MCC}| + c_2 \cdot |1 - \mathrm{DI}|$ for each parameter combination, and the resulting Pareto-optimal points let a user choose a fairness-accuracy trade-off. The experiments report that on the BAF, VIF, and CCF fraud datasets, the chosen balance yields DI ratios near 1 for logistic regression, random forest, SVM, naive Bayes, and Learning Fair Representations, with MCC values that improve on the unmodified baseline.

Load-bearing premise

The load-bearing premise is that the balance parameters selected on a 5,000-row tuning sample (60% train) remain near-optimal for the model on a test set with the original distribution; the paper's own Table 9 shows the transfer is imperfect, with logistic regression's DI ratio dropping from 1.001 on the tuning partition to 0.231 on the larger test collection.

Editorial extensions

If this is right

  • A practitioner with a doubly imbalanced fraud dataset can improve group fairness by resampling the training data only, without changing the classifier or adding a separate debiasing algorithm.
  • The same search can be applied together with methods like Learning Fair Representations, so it acts as a pre-processing layer rather than a replacement for existing debiasing.
  • The Pareto front gives operators a principled way to choose a fairness-accuracy trade-off that satisfies a legal or ethical threshold such as a DI ratio of at least 0.8.
  • Because the three parameters are interpretable, the chosen balance can be audited and reported in plain terms, such as the unprivileged group share and the fraud label share in the training sample.
  • The method also covers singly imbalanced datasets as a boundary case where at least one parameter is fixed at 0, which is the common scenario in many data collections.

Reading between the lines

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

  • Because the optimal parameters differ across classifiers, the search likely compensates for each model's inductive bias; a testable extension is whether a single balance exists that serves an ensemble of models.
  • The same three-parameter search could be applied with other fairness metrics, such as equalized odds, by replacing the DI-ratio loss in the combined objective.
  • The paper evaluates fraud data only, but the method's promise extends to any doubly imbalanced binary classification task, such as credit approval or recidivism prediction, where both the sensitive attribute and the outcome are unevenly distributed.
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

3 major / 5 minor

Summary. The paper addresses fairness in doubly imbalanced datasets, where both the label distribution and the sensitive-attribute group distribution are skewed. It proposes a sampling scheme parameterized by (alpha, beta, gamma) that controls the privilege-group balance, the label balance, and the favourable-rate ratio between groups, followed by a two-level grid search to minimize a combined loss of DI-ratio and MCC. Experiments are carried out on three fraud datasets with five classifiers (LR, NB, RF, SVM, and LFR). The manuscript reports near-ideal DI ratios and improved MCC on the tuning partitions, and also includes results on larger held-out test collections.

Significance. The problem is well motivated: most fairness mitigation methods fail under double imbalance, and a model-agnostic, interpretable sampling schedule would be practically useful. The exploratory analysis in Section 4 is a clear contribution, and the (alpha, beta, gamma) parameterization is transparent. The paper is less convincing as an empirical validation, because the headline numbers in Table 8 are selected on the same partitions used for evaluation, and the independent larger test in Table 9 shows substantial deterioration for most models. The authors do not provide code or data, which limits reproducibility. The central idea is defensible and the manuscript can be strengthened, but the current evidence is not sufficient to support the claim of fine-tuned, transferable optimal balancing.

major comments (3)
  1. [§6.2–6.3, Tables 7–8] The near-ideal DI ratios in Table 8 are fitted, not predictive. In Section 6.3, the two-level grid search selects (alpha, beta, gamma) and the classification threshold (the latter per Section 6.2) by minimizing the combined loss in Eq. (11) on the level-0 and level-1 test partitions, and Table 8 reports results on those same partitions. The only independent check, Table 9, shows that these configurations do not transfer to the larger test collection: for BAF, LR level-0 DI drops from 1.001 to 0.231, RF from 0.995 to 0.535, SVM from 0.990 to 0.544, and LFR from 0.960 to 0.769. The paper should use a separate validation partition for model selection and report final metrics on an untouched test set; without this, the optimality claim is not established.
  2. [§6.3, Table 9] The statement that metric values 'remain similar except that DI Ratio and Combined Loss tend to be somewhat lower and higher respectively' is not supported by the table. NB level-1 degrades from DI 1.007, Recall 0.85, MCC 0.049 to DI 0.545, Recall 0.01, MCC 0.012; LR level-1 DI drops from 0.986 to 0.514. These are not minor or 'expected and acceptable' changes unless a tolerance is explicitly defined. The authors should quantify the acceptable deterioration, report repeated-run variability, and avoid claiming near-consistent performance on the larger test.
  3. [§6.5, Tables 13 and 17] The transfer problem is not specific to the BAF dataset. On CCF level-1, the SVM DI ratio is 0.998 on the tuning partition but 2.681 on the larger test collection; on VIF level-1, the LFR DI ratio is 1.011 on the tuning partition but 0.572 on the larger test. These results are consistent with the grid search overfitting the tuning partition rather than finding a robust balance structure. The authors should either adopt a validation-based selection protocol in all three datasets or substantially soften the claim that the selected parameters are optimal for unseen data.
minor comments (5)
  1. [Figure captions 1–3] The word 'Parento' is misspelled and should be 'Pareto'.
  2. [Equation (2)] The formula for MCC is typeset incorrectly: the denominator should be the square root of the product of the four terms, and the current rendering appears as a single fraction without the square root.
  3. [§5.1, Equations (5)–(8)] The example and surrounding text use theta in place of gamma (e.g., 'parameters alpha, beta, theta = (0.5, 0.8, 0.4)' and the iteration over '(alpha, beta, theta)'), but only alpha, beta, and gamma are defined; this should be made consistent.
  4. [General] No code or data repository is provided, making the experiments difficult to reproduce; the authors should include an artifact link or explicit instructions for reproducing the grid search.
  5. [§6.3] The paper would be stronger with a comparison to simple fixed-ratio resampling baselines (e.g., random undersampling or oversampling to a 50/50 label balance, with and without privilege balancing) to isolate the contribution of the grid-search component.

Circularity Check

2 steps flagged · score 4.0 of 10

Table 8's near-ideal DI ratios are selected in-sample by the grid-search objective; the independent larger test (Table 9) contradicts the headline fairness gains for most BAF models.

  1. fitted input called prediction [Section 5.2 (Algorithm 1) and Section 6.3 (Tables 7-8)]
    "MODEL.evaluate(D' validate, D' test); return(MODEL.MCC LOSS, MODEL.DI RATIO LOSS, MODEL.COMBINED LOSS); ... The fairness and classification accuracy performance obtained when the training dataset is balanced according to the optimal parameters is given in Table 8."

    The grid search selects alpha, beta, gamma by minimizing COMBINED LOSS = c1*MCC_LOSS + c2*DI_RATIO_LOSS, and the losses are computed on the level-0/level-1 test partitions described in Section 6.3. Table 8 then reports the DI Ratio, MCC, and Combined Loss values on exactly those same test partitions. The near-ideal DI ratios (e.g., 1.001 for LR, 0.995 for RF) are therefore the optimized objective values on the tuning partitions, not out-of-sample predictions. The reported 'improvements' are fitted by construction.

  2. fitted input called prediction [Section 6.2 (threshold optimization) with Section 6.3 (Table 8)]
    "Since such classifiers output probabilities, we need to determine a threshold value for mapping these probabilities to class label predictions. This parameter is also optimized within the grid search."

    DI Ratio and MCC depend directly on the classification threshold. The paper optimizes this threshold inside the same grid search that evaluates on the level-0/level-1 test partitions, and Table 8 reports DI and MCC computed with this co-fitted threshold on those same partitions. Thus the fairness improvements in Table 8 are partly artifacts of threshold tuning on the evaluation set, rather than independent measurements of the method's behavior.

full rationale

The paper's central empirical evidence for 'significant improvements in model fairness (DI Ratio)' is Table 8, which reports performance on the same test partitions used by the two-level grid search to minimize |1 - DI| and |1 - MCC|. Those values are selected in-sample, so the near-perfect DI ratios are fitted outcomes rather than predicted outcomes. The paper does provide an independent larger test collection (Table 9 for BAF; Tables 13 and 17 for VIF and CCF) that is not used in parameter selection, which gives genuine non-circular evidence. However, on the BAF dataset this independent test largely fails to support the headline claim: LR Level-0 DI drops from 1.001 to 0.231, RF from 0.995 to 0.535, and SVM from 0.990 to 0.544, all outside the acceptable [0.8, 1.2] band. The paper itself acknowledges the deterioration, calling it 'expected and acceptable,' but this does not remove the circularity burden from Table 8. No self-citation chains or imported uniqueness theorems are present; the circularity is confined to the in-sample selection/reporting of the headline metrics. Score 4 reflects partial circularity with some independent grounding.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim rests on three grid-searched interpolation parameters (alpha, beta, gamma), a hand-set threshold, and choices of sample size and loss weights. The domain assumptions are that DI ratio is the right fairness metric, MCC is the right accuracy metric, and the six search-space restrictions in Equation 3 define a sensible space. No new entities are introduced; the doubly imbalanced label is a framing term, not an entity.

free parameters (6)
  • alpha (group balance weight) = e.g., 0.54 for LR Level 1
    Interpolation weight for the privileged/unprivileged group rate; selected by grid search on the validation partition (Table 7).
  • beta (label balance weight) = e.g., 0.00 for LR Level 1
    Interpolation weight for the favourable/unfavourable label rate; selected by grid search on the validation partition (Table 7).
  • gamma (favourable-rate ratio weight) = e.g., 1.00 for LR Level 1
    Interpolation weight for the ratio of favourable-label rates between privileged and unprivileged groups; selected by grid search (Table 7).
  • classification threshold = not reported per model
    The probability-to-label threshold is optimized within the grid search (Section 6.2), adding another fitted component to the reported metrics.
  • c1, c2 (combined loss weights) = 1, 1
    Weights in the combined loss (Equation 11) are set by hand to 1, with no sensitivity analysis; the Pareto front depends on this choice.
  • training sample size = 5000
    The grid search is run on a 5,000-row subset chosen for efficiency (Section 6.3), which may not represent the original distribution.
assumptions (5)
  • domain assumption DI ratio is an appropriate and sufficient fairness metric.
    The entire optimization uses DI ratio as the sole fairness objective (Equations 1, 9), without considering other fairness definitions.
  • domain assumption MCC is the appropriate accuracy metric for imbalanced binary classification.
    MCC is used as the classification objective (Equations 2, 10); the paper justifies this choice but does not test sensitivity to other metrics.
  • domain assumption The six restrictions in Equation 3 define a sensible search space.
    The restrictions prevent role reversal between privileged and unprivileged groups and limit how much imbalance can be increased; these are policy choices, not mathematical necessities.
  • domain assumption The sensitive attribute splits (old vs young by age >= 50; female vs male) are meaningful privilege proxies.
    The experiments assume these binary splits capture the relevant privilege structure in each fraud dataset.
  • domain assumption The LFR implementation in AIF360 is reliable and representative of debiasing methods.
    The exploratory analysis uses LFR as the sole representative debiasing method, and the paper generalizes from its failure to other debiasing methods.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fair for a few: Improving Fairness in Doubly Imbalanced Datasets." pith.science (2026). https://pith.science/paper/VRVYAOPI

@misc{pith2026250614306,
  author       = {Pith},
  title        = {Pith review of: Fair for a few: Improving Fairness in Doubly Imbalanced Datasets},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VRVYAOPI}},
  note         = {Machine review of arXiv:2506.14306}
}
read the original abstract

Fairness has been identified as an important aspect of Machine Learning and Artificial Intelligence solutions for decision making. Recent literature offers a variety of approaches for debiasing, however many of them fall short when the data collection is imbalanced. In this paper, we focus on a particular case, fairness in doubly imbalanced datasets, such that the data collection is imbalanced both for the label and the groups in the sensitive attribute. Firstly, we present an exploratory analysis to illustrate limitations in debiasing on a doubly imbalanced dataset. Then, a multi-criteria based solution is proposed for finding the most suitable sampling and distribution for label and sensitive attribute, in terms of fairness and classification accuracy

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 31 canonical work pages

  1. [1]

    arXiv (2015)

    Feldman, M., Friedler, S., Moeller, J., Scheidegger, C., Venkatasubramanian, S.: Certifying and Removing Disparate Impact. arXiv (2015)

    work page 2015

  2. [2]

    AAAI 32(1) (2018) https://doi.org/10.1609/aaai.v32i1.11296

    Grgi´ c-Hlaˇ ca, N., Zafar, M.B., Gummadi, K.P., Weller, A.: Beyond distributive fairness in algorithmic decision making: Feature selection for procedurally fair learning. AAAI 32(1) (2018) https://doi.org/10.1609/aaai.v32i1.11296

    work page doi:10.1609/aaai.v32i1.11296 2018

  3. [3]

    Human Perceptions of Fairness in Algorithmic Decision Making: A Case Study of Criminal Risk Prediction

    Grgi´ c-Hlaˇ ca, N., Redmiles, E.M., Gummadi, K.P., Weller, A.: Human Percep- tions of Fairness in Algorithmic Decision Making: A Case Study of Criminal Risk Prediction. arXiv (2018). https://doi.org/10.48550/arXiv.1802.09548

    work page Pith review arXiv doi:10.48550/arxiv.1802.09548 2018

  4. [4]

    IEEE Access 10, 120850–120865 (2022) 28 T able 17: Performance Results: Analysis on the Larger Test Collec- tion of CCF dataset

    Gittens, A., Yener, B., Yung, M.: An adversarial perspective on accuracy, robust- ness, fairness, and privacy: multilateral-tradeoffs in trustworthy ml. IEEE Access 10, 120850–120865 (2022) 28 T able 17: Performance Results: Analysis on the Larger Test Collec- tion of CCF dataset. (Precision, Recall and F1 values are obtained for the Fraud label.) Grid Se...

    work page 2022

  5. [5]

    Manage- ment Science 65(7), 2966–2981 (2019) https://doi.org/10.1287/mnsc.2018.3093 https://doi.org/10.1287/mnsc.2018.3093

    Lambrecht, A., Tucker, C.: Algorithmic bias? an empirical study of appar- ent gender-based discrimination in the display of stem career ads. Manage- ment Science 65(7), 2966–2981 (2019) https://doi.org/10.1287/mnsc.2018.3093 https://doi.org/10.1287/mnsc.2018.3093

    arXiv 2019

  6. [6]

    MIT Press, ??? (2023)

    Barocas, S., Hardt, M., Narayanan, A.: Fairness and Machine Learning. MIT Press, ??? (2023)

    work page 2023

  7. [7]

    ACM Comput

    Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., Galstyan, A.: A survey on bias and fairness in machine learning. ACM Comput. Surv. 54(6), 1–35 (2022) https://doi.org/10.1145/3457607

    doi:10.1145/3457607 2022

  8. [9]

    In: Proceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering, pp

    Biswas, S., Rajan, H.: Fair preprocessing: Towards understanding compositional fairness of data transformers in machine learning pipeline. In: Proceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering, pp. 981–993. ACM, Athens Greece (2021). https://doi.org/10.1145/346826...

    arXiv 2021

Show all 41 references
  1. [10]

    In: 2009 2nd Inter- national Conference on Computer, Control and Communication, pp

    Kamiran, F., Calders, T.: Classifying without discriminating. In: 2009 2nd Inter- national Conference on Computer, Control and Communication, pp. 1–6. IEEE, 29 (a) Pareto front with LR method with Grid Search Level 0 and Level 1 (b) Pareto front with RF method with Grid Search...

    2009

  2. [11]

    arXiv (2021)

    Richardson, B., Gilbert, J.E.: A Framework for Fairness: A Systematic Review of Existing Fair AI Solutions. arXiv (2021)

    2021

  3. [12]

    Yang, K., Huang, B., Stoyanovich, J., Schelter, S.: Fairness-aware instrumentation 30 of preprocessing pipelines for machine learning (2020)

    2020

  4. [13]

    IEEE Access (2024)

    Iurada, L., Bucci, S., Hospedales, T.M., Tommasi, T.: Fairness meets cross- domain learning: A benchmark of models and metrics. IEEE Access (2024)

    2024

  5. [14]

    In: 2009 IEEE International Conference on Data Mining Workshops, pp

    Calders, T., Kamiran, F., Pechenizkiy, M.: Building classifiers with independency constraints. In: 2009 IEEE International Conference on Data Mining Workshops, pp. 13–18. IEEE, Miami, FL, USA (2009). https://doi.org/10.1109/ICDMW. 2009.83

    2009 doi

  6. [15]

    In: Proceedings of the 28th ACM International Conference on Information and Knowledge Management, pp

    Iosifidis, V., Ntoutsi, E.: Adafair: Cumulative fairness adaptive boosting. In: Proceedings of the 28th ACM International Conference on Information and Knowledge Management, pp. 781–790. ACM, Beijing China (2019). https://doi. org/10.1145/3357384.3357974

    2019

  7. [16]

    In: 2021 International Conference on Applied Artificial Intelligence (ICAPAI), pp

    Kenfack, P.J., Khan, A.M., Kazmi, S.M.A., Hussain, R., Oracevic, A., Khattak, A.M.: Impact of model ensemble on the fairness of classifiers in machine learning. In: 2021 International Conference on Applied Artificial Intelligence (ICAPAI), pp. 1–6. IEEE, Halden, Norway (2021)....

    2021

  8. [17]

    arXiv (2019)

    Saleiro, P., Kuester, B., Hinkson, L., London, J., Stevens, A., Anisfeld, A., Rodolfa, K.T., Ghani, R.: Aequitas: A Bias and Fairness Audit Toolkit. arXiv (2019)

    2019

  9. [18]

    arXiv (2018)

    Bellamy, R.K.E., Dey, K., Hind, M., Hoffman, S.C., Houde, S., Kannan, K., Lohia, P., Martino, J., Mehta, S., Mojsilovic, A., Nagar, S., Ramamurthy, K.N., Richards, J., Saha, D., Sattigeri, P., Singh, M., Varshney, K.R., Zhang, Y.: AI Fairness 360: An Extensible Toolkit for Det...

    2018

  10. [19]

    Journal of artificial intelligence research 16, 321–357 (2002)

    Chawla, N.V., Bowyer, K.W., Hall, L.O., Kegelmeyer, W.P.: Smote: synthetic minority over-sampling technique. Journal of artificial intelligence research 16, 321–357 (2002)

    2002

  11. [20]

    Journal of Biomedical Informatics 107, 103465 (2020)

    Xu, Z., Shen, D., Nie, T., Kou, Y.: A hybrid sampling algorithm combining m- smote and enn based on random forest for medical imbalanced data. Journal of Biomedical Informatics 107, 103465 (2020)

    2020

  12. [21]

    In: International Conference on Intelligent Computing, pp

    Han, H., Wang, W.-Y., Mao, B.-H.: Borderline-smote: a new over-sampling method in imbalanced data sets learning. In: International Conference on Intelligent Computing, pp. 878–887 (2005). Springer

    2005

  13. [22]

    In: 2020 IEEE 21st International Conference on Information Reuse and Integration for Data Science (IRI), pp

    Yilmaz, I., Masum, R., Siraj, A.: Addressing imbalanced data problem with gener- ative adversarial network for intrusion detection. In: 2020 IEEE 21st International Conference on Information Reuse and Integration for Data Science (IRI), pp. 25–30 (2020). IEEE 31

    2020

  14. [23]

    In: International Conference on Medical Image Computing and Computer-assisted Intervention, pp

    Marrakchi, Y., Makansi, O., Brox, T.: Fighting class imbalance with con- trastive learning. In: International Conference on Medical Image Computing and Computer-assisted Intervention, pp. 466–476 (2021). Springer

    2021

  15. [24]

    IEEE Access 9, 109960–109975 (2021)

    Khushi, M., Shaukat, K., Alam, T.M., Hameed, I.A., Uddin, S., Luo, S., Yang, X., Reyes, M.C.: A comparative performance analysis of data resampling methods on imbalance medical data. IEEE Access 9, 109960–109975 (2021)

    2021

  16. [25]

    IEEE Access 7, 93010–93022 (2019)

    Makki, S., Assaghir, Z., Taher, Y., Haque, R., Hacid, M.-S., Zeineddine, H.: An experimental study with imbalanced classification approaches for credit card fraud detection. IEEE Access 7, 93010–93022 (2019)

    2019

  17. [26]

    IEEE Transactions on Biometrics, Behavior, and Identity Science 5(2), 244–254 (2022)

    Nagpal, S., Singh, M., Singh, R., Vatsa, M.: Detox loss: Fairness constraints for learning with imbalanced data. IEEE Transactions on Biometrics, Behavior, and Identity Science 5(2), 244–254 (2022)

    2022

  18. [27]

    IEEE Access11, 56691–56702 (2023)

    Lavalle, A., Mate, A., Trujillo, J., Teruel, M.A.: A data analytics methodology to visually analyze the impact of bias and rebalancing. IEEE Access11, 56691–56702 (2023)

    2023

  19. [28]

    Expert Systems with Applications 228, 120323 (2023)

    Sha, L., Gaˇ sevi´ c, D., Chen, G.: Lessons from debiasing data for fair and accurate predictive modeling in education. Expert Systems with Applications 228, 120323 (2023)

    2023

  20. [29]

    In: 2015 IEEE Symposium Series on Computational Intelligence, pp

    Dal Pozzolo, A., Caelen, O., Johnson, R.A., Bontempi, G.: Calibrating probabil- ity with undersampling for unbalanced classification. In: 2015 IEEE Symposium Series on Computational Intelligence, pp. 159–166 (2015). IEEE

    2015

  21. [30]

    Advances in Neural Information Processing Systems (2022)

    Jesus, S., Pombal, J., Alves, D., Cruz, A., Saleiro, P., Ribeiro, R.P., Gama, J., Bizarro, P.: Turning the tables: Biased, imbalanced, dynamic tabular datasets for ml evaluation. Advances in Neural Information Processing Systems (2022)

    2022

  22. [31]

    In: Proceedings of the 35th International Conference on Machine Learning, pp

    Kearns, M., Neel, S., Roth, A., Wu, Z.S.: Preventing fairness gerrymander- ing: Auditing and learning for subgroup fairness. In: Proceedings of the 35th International Conference on Machine Learning, pp. 2564–2572. PMLR, ??? (2018)

    2018

  23. [32]

    In: Proceedings of the Conference on Fairness, Accountability, and Transparency

    Kearns, M., Neel, S., Roth, A., Wu, Z.S.: An empirical study of rich subgroup fairness for machine learning. In: Proceedings of the Conference on Fairness, Accountability, and Transparency. F AT* ’19, pp. 100–109. Association for Com- puting Machinery, New York, NY, USA (2019)...

    2019 doi

  24. [33]

    In: Proceedings of the Conference on Fairness, Accountability, and Transparency

    Celis, L.E., Huang, L., Keswani, V., Vishnoi, N.K.: Classification with fairness constraints: A meta-algorithm with provable guarantees. In: Proceedings of the Conference on Fairness, Accountability, and Transparency. F AT* ’19, pp. 319–

  25. [34]

    arXiv (2023)

    Pombal, J., Saleiro, P., Figueiredo, M.A.T., Bizarro, P.: Fairness-Aware Data Val- uation for Supervised Learning. arXiv (2023). https://doi.org/10.48550/arXiv. 2303.16963

    2023 doi arXiv

  26. [35]

    Information sciences 501, 118–135 (2019)

    Douzas, G., Bacao, F.: Geometric smote a geometrically enhanced drop-in replacement for smote. Information sciences 501, 118–135 (2019)

    2019

  27. [36]

    Expert Systems with Applications 178, 115011 (2021)

    Moniz, N., Cerqueira, V.: Automated imbalanced classification via meta-learning. Expert Systems with Applications 178, 115011 (2021)

    2021

  28. [37]

    BMC genomics 21, 1–13 (2020)

    Chicco, D., Jurman, G.: The advantages of the matthews correlation coeffi- cient (mcc) over f1 score and accuracy in binary classification evaluation. BMC genomics 21, 1–13 (2020)

    2020

  29. [38]

    Pattern Recognition Letters 136, 71–80 (2020)

    Zhu, Q.: On the performance of matthews correlation coefficient (mcc) for imbalanced dataset. Pattern Recognition Letters 136, 71–80 (2020)

    2020

  30. [39]

    In: Proceedings of the 30th International Conference on Machine Learning, pp

    Zemel, R., Wu, Y., Swersky, K., Pitassi, T., Dwork, C.: Learning fair representa- tions. In: Proceedings of the 30th International Conference on Machine Learning, pp. 325–333. PMLR, ??? (2013)

    2013

  31. [40]

    Neurocomputing 415, 295–316 (2020) https://doi

    Yang, L., Shami, A.: On hyperparameter optimization of machine learning algo- rithms: Theory and practice. Neurocomputing 415, 295–316 (2020) https://doi. org/10.1016/j.neucom.2020.07.061

    2020 doi

  32. [41]

    In: Esparza, J., Majumdar, R

    Legriel, J., Le Guernic, C., Cotton, S., Maler, O.: Approximating the pareto front of multi-criteria optimization problems. In: Esparza, J., Majumdar, R. (eds.) Tools and Algorithms for the Construction and Analysis of Systems, pp. 69–83. Springer, Berlin, Heidelberg (2010) 33

    2010

  33. [328]

    https: 32 //doi.org/10.1145/3287560.3287586

    Association for Computing Machinery, New York, NY, USA (2019). https: 32 //doi.org/10.1145/3287560.3287586

    2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.