Pith. sign in

REVIEW 3 major objections 4 minor 43 references

Neighbor displacement-based enhanced synthetic oversampling for multiclass imbalanced data

T0 review · 3 major / 4 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 This paper claims that relocating noisy minority-class points toward their class centroids before random oversampling produces a multiclass resampler that outperforms 14 baselines on average G-mean across 20 datasets and nine classifiers.

desk verdict Interesting displacement-before-oversampling idea, but the evaluation resamples before the split, so the headline results don't measure generalization. read the letter →

arxiv 2501.04099 v1 pith:5INJEH7R submitted 2025-01-07 cs.LG

I Made Putrama , Peter Martinek This is my paper · ORCID
classification cs.LG
keywords multiclassimbalanceoversamplingrandomneighbordisplacementk-nearestneighborsG-meanFriedman-Nemenyitestimbalancedclassification
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 proposes NDESO, a hybrid resampling method for multiclass imbalanced classification. Its idea is to clean before balancing: find points whose k nearest neighbors are mostly from other classes, move them closer to their own class centroid by one average-neighbor-distance, and only then apply random oversampling. This preserves every point's class label while reducing overlap, which the authors argue avoids the information loss of undersampling and the noise amplification of SMOTE-style interpolation in sparse data. On 20 real-world datasets across nine classifiers, they report that NDESO achieves the highest average G-mean and the lowest mean rank among 15 resamplers, with the largest gains on the most extreme-imbalance datasets. If the comparison holds up, this is a low-cost, label-preserving preprocessing step that could be dropped into existing imbalanced-learning pipelines.

What carries the argument

The load-bearing mechanism is the NDE displacement rule. For a point $x_i$ whose $k$-nearest-neighbor neighborhood is majority other-class, with class centroid $r_{c_i}$, normalized direction $\vec{S}_v = (r_{c_i}-x_i)/S_i$, and $\phi_i$ the average distance from $x_i$ to its $k$ neighbors, the point is moved to $x'_i = r_{c_i} - \vec{S}_v\,\phi_i$. The rule compresses overlapping regions by pulling disputed points toward their own class center while keeping the original labels, and it requires only pairwise distances and a centroid, which is why it can run on very sparse classes. Random oversampling then balances the cleaned distribution.

What would settle it

Re-run the same comparison with the resampling pipeline embedded inside each cross-validation fold, fitting the NDE displacement and random oversampling only on the training portion, and compare the resulting G-mean ranks. NDESO's claimed edge would be falsified if its mean rank no longer leads, or if its average G-mean advantage over SMOTE-ENN and SMOTE-CDNN shrinks below the Nemenyi critical difference.

Watch

Extended reading notes

Core claim

The central claim is that noisy minority-class points should be repositioned rather than removed or relabeled. A point counts as noisy when more than half of its k nearest neighbors belong to different classes; NDE then computes the average distance from that point to those neighbors and shifts the point along the unit vector toward its own class centroid by exactly that distance. Random oversampling then duplicates minority points until class sizes match the majority. In the authors' experiments this two-stage procedure outperforms 14 baseline resamplers, including SMOTE variants, ADASYN, and SMOTE-CDNN, in average G-mean and in Friedman/Nemenyi mean rank, and it runs successfully on sparse datasets where several baselines fail because a minority class has too few members.

Load-bearing premise

The experiments assume that resampling the whole dataset before the training/test split gives an unbiased estimate of generalization; if synthetic copies of test points can leak into training, the reported G-mean advantage may not reflect performance on new data.

Editorial extensions

If this is right

  • Datasets with an extreme imbalance ratio, up to 853:1 in the paper, are where NDESO shows its largest gains, suggesting the cleaning step matters most when minority classes are tiny.
  • Because NDESO displaces rather than deletes or relabels points, no minority class is erased by preprocessing, in contrast to undersampling or edited-neighbor methods that can collapse a class to zero members.
  • NDESO runs on sparse datasets where SMOTE-family samplers fail with neighbor-count errors, which broadens the set of multiclass problems a single resampler can handle.
  • The method's extra cost over plain random oversampling is the pairwise-distance computation of NDE, so the reported accuracy gain is obtained without a generative model.

Reading between the lines

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

  • The main risk to the ranking is evaluation leakage: the protocol resamples before splitting, so a leak-free re-run with the resampler fitted inside each fold is the decisive test of whether NDESO's G-mean advantage is real.
  • NDE could be treated as a generic denoising front end: composing it with oversamplers other than ROS, such as SMOTE or ADASYN, might inherit the cleaning benefit while adding synthetic diversity; the paper's own NDE+SMOTE experiments suggest the displacement, not the final sampler, drives much of the gain.
  • Because the displacement rule only needs a distance metric, the same cleaning idea could transfer to image, spatial, or graph data once a pairwise distance is defined; the paper names big-data and non-tabular resampling as open directions.
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

3 major / 4 minor

Summary. The paper proposes NDESO (Neighbor Displacement-based Enhanced Synthetic Oversampling), a hybrid resampler that first identifies points whose k-nearest neighbors are mostly from other classes, displaces those points toward their own class centroid, and then applies random oversampling to balance the class distribution. The authors evaluate NDESO against 14 baseline resamplers on 20 multiclass datasets with 9 classifiers, using G-mean and the Friedman/Nemenyi tests, and claim that NDESO achieves the highest average G-mean and the lowest mean rank. The core contribution is a displacement-based alternative to cleaning methods that delete noisy minority instances.

Significance. The idea of repositioning noisy minority points rather than deleting them is a plausible and potentially useful contribution to imbalanced multiclass resampling. The paper also provides a clear algorithmic description, a GitHub repository for reproducibility, and a broad comparison across datasets and classifiers. However, the main empirical claim is not supported by the experimental design as reported, because the evaluation protocol leaks test information into training. If the experiments were correctly re-run with the resampler fitted only on training folds, the rankings could change substantially. The current manuscript therefore does not establish the claimed superiority of NDESO.

major comments (3)
  1. [Section 5.2] The testing procedure resamples the full dataset before splitting: "A given dataset is tested using a resampling method to balance the class distributions. After resampling, the dataset is then partitioned into 80% training and 20% testing subsets using cross-validation." In NDESO, Algorithm 1 computes pairwise distances over the entire dataset, determines displaceable points using labels from the full dataset, and then random oversampling duplicates minority instances before the split. This means test instances can be repositioned or their synthetic copies can be placed into the training set, while the original test instances remain in the test set. For classifiers such as k-NN, SVC, and tree ensembles, this inflates the reported G-mean values and invalidates all ranks used in the Friedman and Nemenyi tests. The resampler must be fitted only within each training fold. This is a load-bearing flaw: the abstract's claim that NDESO 'outperforms its competitors regarding average G-mean score and achieves the lowest statistical mean rank' is not supported by the current protocol.
  2. [Section 5.3.2] The Nemenyi test results as described contradict the claim of statistical superiority. The paper states that 'no significant differences exist among ... Borderline-SMOTE, SMOTE-CDNN, SMOTE-ENN, and NDESO' and then immediately asserts that 'with its lowest average mean rank of 1.85 ... it signifies better performance.' Being in the same non-significant block as three other methods means the observed rank advantage over those methods is not statistically significant. The wording must be corrected to report the actual inference: NDESO is statistically tied with those three methods and significantly better than the remaining ones. As written, the discussion overstates the statistical evidence.
  3. [Tables 3 and 5] Many baseline methods have missing entries due to resampling failures (Table 2 lists 9 datasets for KMeans-SMOTE, 6 for SMOTE-ENN/ENN/ECDNN, etc.), yet average G-mean values and mean ranks are reported for all methods in Tables 3 and 5. The paper does not state how averages and ranks are computed in the presence of these missing values. If a method is ranked only on the datasets where it succeeded, different methods are compared on different subsets of datasets, which biases the comparison in favor of methods that always succeed (such as NDESO). The Friedman test also requires complete block data, so the reported p-value of 6.08e-20 is not interpretable unless the analysis was confined to the subset of datasets where all methods ran successfully or a valid incomplete-data method was used. The authors should either restrict all comparisons to the common subset of complete datasets, or use a paired analysis that explicitly accounts for missing values.
minor comments (4)
  1. [Section 4.1, Eq. (5)] The displacement formula x'_i = r_ci - S_v * phi_i moves the point beyond the centroid when the average neighbor distance phi_i exceeds the distance from x_i to the centroid. No condition is given to prevent the point from being relocated to the opposite side of the class center. The paper should specify the intended behavior or add clipping.
  2. [Algorithm 1, lines 14-17] The pseudocode sets R = X[y=cls] for each class and then refers to R[i] as the centroid for point i. Since R is a list of class-specific arrays, this indexing is not well-defined; the centroid for a point should be r_{c_i} as in Eq. (2). Please correct the notation.
  3. [Section 5.2] The number of cross-validation folds is computed as nsplits = min(5, min(ytrain)), but ytrain is defined only after the resampling step. Since resampling balances the classes, min(ytrain) will typically be the size of the majority class and will not reflect the original minority class size. Clarify the intended definition and the actual CV procedure used.
  4. [General presentation] Minor typographical issues include 'eucledian' in Algorithm 1 and the inconsistent formatting of class counts in Table 1 (e.g., the segment row has an unclosed bracket). These are cosmetic but should be cleaned up.

Circularity Check

1 steps flagged · score 6.0 of 10

NDESO's superiority claim rests on a circular evaluation protocol: Section 5.2 resamples the full dataset before the train/test split, so test points and their synthetic duplicates leak into training.

  1. other [Section 5.2 ('Testing procedure'); Algorithm 1 (NDESO/NDE), lines 2, 20, 24-26]
    "Section 5.2: "A given dataset is tested using a resampling method to balance the class distributions. After resampling, the dataset is then partitioned into 80% training and 20% testing subsets using cross-validation." Algorithm 1: "d = cdist(X, X)" ... "X[i] = rci - Sv * phi_i" ... "X, y= NDE(X, y, cdist, k)" and "RANDOM OVER(X, y)"."

    NDESO is applied to the full dataset X before any split: NDE computes pairwise distances over all points, displaces points using Eq. (5), and random oversampling duplicates minority samples. Only after this does Section 5.2 partition the resampled data into train/test. Consequently, the 20% 'test' instances are part of the input used to construct the training set: their positions are changed and exact duplicates of them are added to the training pool. Classifiers such as k-NN, SVC, and tree ensembles can then effectively memorize the test points, and the reported G-mean and Friedman/Nemenyi ranks are computed on an in-sample evaluation.

full rationale

The NDE displacement rule and the ROS duplication are defined directly on the data; no parameter is fitted to the G-mean and no self-citation is load-bearing. However, the experimental protocol in Section 5.2 resamples the entire dataset before the 80/20 partition. Algorithm 1's cdist(X,X), displacement update, and random oversampling therefore operate on the same X that later supplies the test fold, and the random oversampling can place exact duplicates of test instances into the training set. Because k-NN, SVC, and tree ensembles are sensitive to near-duplicates, the reported average G-mean of 0.9362 and mean rank of 2.88 are in-sample numbers, not independent predictions. This makes the headline empirical claim partially circular: the 'test' performance is produced by construction from the test data themselves. The algorithm itself is not a disguised re-derivation of the benchmark, and there is no self-citation chain, so the circularity is confined to the evaluation protocol.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The algorithm itself is simple and free of fitted coefficients, but it rests on a noise model, a displacement assumption, an unstated geometric condition, and a flawed evaluation assumption. The noise model equates label disagreement among k neighbors with noisiness; the displacement formula only moves points toward the centroid when the average neighbor distance is smaller than the distance to the centroid; and Section 5.2 resamples before the train/test split. These assumptions are load-bearing for the reported results.

free parameters (1)
  • k (number of neighbors) = 5
    Chosen as default; Table 6 reports k=5 gives the highest average G-mean, so the headline comparison uses the best-performing k on the same benchmark data.
assumptions (5)
  • domain assumption A data point whose k-nearest neighbors are mostly from other classes is noisy and should be repositioned toward its own class centroid.
    Section 4.1, Eq. (1): displaceable points are defined by A < B, i.e., more neighbors from other classes than from the same class.
  • domain assumption Moving those points closer to their centroid and then random oversampling improves class separation and downstream G-mean.
    Section 4.2 pairs NDE with random oversampling; no proof is given that this preserves the data distribution or reduces overlap after resampling.
  • ad hoc to paper Eq. (5) moves each displaced point closer to its centroid; this requires the average neighbor distance phi_i to be smaller than the point's distance to the centroid.
    The formula x'_i = r_ci - S_v * phi_i places the new point on the far side of the centroid at distance phi_i; the paper states 'displace xi closer to its centroid' without stating or checking the phi_i < ||x_i - r_ci|| condition.
  • ad hoc to paper Resampling the full dataset before the train/test split yields valid generalization estimates.
    Section 5.2 applies the resampler before the 80/20 split; standard practice resamples only training folds, so this assumption is load-bearing and likely false.
  • domain assumption The binary G-mean definition in Eq. (7) can be applied to multiclass problems without further specification.
    Section 4.3.1 defines G-mean for sensitivity/specificity but does not explain how it is averaged across more than two classes in the experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Neighbor displacement-based enhanced synthetic oversampling for multiclass imbalanced data." pith.science (2026). https://pith.science/paper/5INJEH7R

@misc{pith2026250104099,
  author       = {Pith},
  title        = {Pith review of: Neighbor displacement-based enhanced synthetic oversampling for multiclass imbalanced data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5INJEH7R}},
  note         = {Machine review of arXiv:2501.04099}
}
read the original abstract

Imbalanced multiclass datasets pose challenges for machine learning algorithms. These datasets often contain minority classes that are important for accurate prediction. Existing methods still suffer from sparse data and may not accurately represent the original data patterns, leading to noise and poor model performance. A hybrid method called Neighbor Displacement-based Enhanced Synthetic Oversampling (NDESO) is proposed in this paper. This approach uses a displacement strategy for noisy data points, computing the average distance to their neighbors and moving them closer to their centroids. Random oversampling is then performed to achieve dataset balance. Extensive evaluations compare 14 alternatives on nine classifiers across synthetic and 20 real-world datasets with varying imbalance ratios. The results show that our method outperforms its competitors regarding average G-mean score and achieves the lowest statistical mean rank. This highlights its superiority and suitability for addressing data imbalance in practical applications.

Figures

Figures reproduced from arXiv: 2501.04099 by the authors.

Figure 1
Figure 1. Visualization of resampling on a sparse multiclass dataset: (a) original dataset; (b) noisy resampled dataset [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Visual illustration of the CDNN algorithm [32] [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The overlapping (before) and cleaned (after) data points [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Visual illustration of displace-able data point identification [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: G-mean scores of our NDE algorithm evaluated across three classifiers for various [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: G-mean scores of our NDE algorithm evaluated across three classifiers for various distance metrics [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Scatter plots of our NDE algorithm compared to other algorithms [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: G-mean scores of our NDE algorithm evaluated across three classifiers for various samplers [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Comparison of average G-mean scores of resampling methods across classifiers and datasets [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Scatter plots of resampling methods applied to the [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: Confusion matrices for resampling methods evaluated using G-mean scores on [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 21 canonical work pages

  1. [1]

    Thippa Reddy, Celestine Iwendi, Ali Kashif Bashir, and Ohyun Jo

    Harshita Patel, Dharmendra Singh Rajput, G. Thippa Reddy, Celestine Iwendi, Ali Kashif Bashir, and Ohyun Jo. A review on classification of imbalanced data for wireless sensor networks. Int. J. Distrib. Sens. Networks, 16(4),

    work page

  2. [2]

    Random CapsNet forest model for imbalanced malware type classification task

    Aykut Çayır, U˘gur Ünal, and Hasan Da˘g. Random CapsNet forest model for imbalanced malware type classification task. Comput. Secur., 102, 2021. ISSN 01674048. doi:10.1016/j.cose.2020.102133

    arXiv 2021

  3. [3]

    Automated imbalanced classification via meta-learning

    Nuno Moniz and Vitor Cerqueira. Automated imbalanced classification via meta-learning. Expert Syst. Appl., 178 (March):115011, 2021. ISSN 09574174. doi:10.1016/j.eswa.2021.115011

    work page arXiv 2021

  4. [4]

    RN-SMOTE: Reduced Noise SMOTE based on DBSCAN for enhancing imbalanced data classification

    Ahmed Arafa, Nawal El-Fishawy, Mohammed Badawy, and Marwa Radad. RN-SMOTE: Reduced Noise SMOTE based on DBSCAN for enhancing imbalanced data classification. J. King Saud Univ. - Comput. Inf. Sci., 34(8): 5059–5074, 2022. ISSN 22131248. doi:10.1016/j.jksuci.2022.06.005

    work page doi:10.1016/j.jksuci.2022.06.005 2022

  5. [5]

    Grouping-based Oversampling in Kernel Space for Imbalanced Data Classification

    Jinjun Ren, Yuping Wang, Yiu ming Cheung, Xiao Zhi Gao, and Xiaofang Guo. Grouping-based Oversampling in Kernel Space for Imbalanced Data Classification. Pattern Recognit., 133:108992, 2023. ISSN 00313203. doi:10.1016/j.patcog.2022.108992

    work page arXiv 2023

  6. [6]

    Self-adaptive oversampling method based on the complexity of minority data in imbalanced datasets classification

    Xinmin Tao, Xinyue Guo, Yujia Zheng, Xiaohan Zhang, and Zhiyu Chen. Self-adaptive oversampling method based on the complexity of minority data in imbalanced datasets classification. Knowledge-Based Syst., 277: 110795, 2023. ISSN 09507051. doi:10.1016/j.knosys.2023.110795

    arXiv 2023

  7. [7]

    Review of resampling techniques for the treatment of imbalanced industrial data classification in equipment condition monitoring

    Yage Yuan, Jianan Wei, Haisong Huang, Weidong Jiao, Jiaxin Wang, and Hualin Chen. Review of resampling techniques for the treatment of imbalanced industrial data classification in equipment condition monitoring. Eng. Appl. Artif. Intell., 126(PB):106911, 2023. ISSN 09521976. doi:10.1016/j.engappai.2023.106911

    work page arXiv 2023

  8. [8]

    Madkour, Hatem M

    Ahmed H. Madkour, Hatem M. Abdelkader, and Amgad M. Mohammed. Dynamic Classification Ensembles for Handling Imbalanced Multiclass Drifted Data Streams. Inf. Sci. (Ny)., 670(April):120555, 2024. ISSN 00200255. doi:10.1016/j.ins.2024.120555

    work page arXiv 2024

Show all 43 references
  1. [9]

    What makes multi-class imbalanced problems difficult? An experimental study

    Mateusz Lango and Jerzy Stefanowski. What makes multi-class imbalanced problems difficult? An experimental study. Expert Syst. Appl., 199(April), 2022. ISSN 09574174. doi:10.1016/j.eswa.2022.116962

    2022

  2. [10]

    Classification of Imbalanced Data Set in Financial Field Based on Combined Algorithm

    Tingting Yu and Yunxiang Huo. Classification of Imbalanced Data Set in Financial Field Based on Combined Algorithm. Mob. Inf. Syst., 2022, 2022. ISSN 1875905X. doi:10.1155/2022/1839204

    2022 doi

  3. [11]

    Deep transfer network with joint distribution adaptation: A new intelligent fault diagnosis framework for industry application

    Te Han, Chao Liu, Wenguang Yang, and Dongxiang Jiang. Deep transfer network with joint distribution adaptation: A new intelligent fault diagnosis framework for industry application. ISA Trans., 97:269–281, 2020. ISSN 00190578. doi:10.1016/j.isatra.2019.08.012

    2020 doi

  4. [12]

    Bajaj, Abhishek D

    Naman S. Bajaj, Abhishek D. Patange, R. Jegadeeshwaran, Sujit S. Pardeshi, Kaushal A. Kulkarni, and Rohan S. Ghatpande. Application of metaheuristic optimization based support vector machine for milling cutter health monitoring. Intell. Syst. with Appl., 18(February):200196, 2...

    2023

  5. [13]

    Imbalanced Sample Selection with Deep Reinforcement Learning for Fault Diagnosis

    Saite Fan, Xinmin Zhang, and Zhihuan Song. Imbalanced Sample Selection with Deep Reinforcement Learning for Fault Diagnosis. IEEE Trans. Ind. Informatics , 18(4):2518–2527, 2022. ISSN 19410050. doi:10.1109/TII.2021.3100284

    2022

  6. [14]

    Systematic review of class imbalance problems in manufac- turing

    Andrea de Giorgio, Gabriele Cola, and Lihui Wang. Systematic review of class imbalance problems in manufac- turing. J. Manuf. Syst., 71(September):620–644, 2023. ISSN 02786125. doi:10.1016/j.jmsy.2023.10.014. 18 Neighbor displacement-based enhanced synthetic oversampling for m...

    2023 doi

  7. [15]

    A broad review on class imbalance learning techniques

    Salim Rezvani and Xizhao Wang. A broad review on class imbalance learning techniques. Appl. Soft Comput., 143:110415, 2023. ISSN 15684946. doi:10.1016/j.asoc.2023.110415

    2023

  8. [16]

    Boosting methods for multi-class imbalanced data classification: an experimental review

    Jafar Tanha, Yousef Abdi, Negin Samadi, Nazila Razzaghi, and Mohammad Asadpour. Boosting methods for multi-class imbalanced data classification: an experimental review. J. Big Data, 7(1), 2020. ISSN 21961115. doi:10.1186/s40537-020-00349-y

    2020 doi

  9. [17]

    A modified real-value negative selection detector-based over- sampling approach for multiclass imbalance problems

    Ming Liu, Minggang Dong, and Chao Jing. A modified real-value negative selection detector-based over- sampling approach for multiclass imbalance problems. Inf. Sci. (Ny)., 556:160–176, 2021. ISSN 00200255. doi:10.1016/j.ins.2020.12.058

    2021 doi

  10. [18]

    M.Nadim Shahariar, and Muhammed J.A

    Nazim Uddin Niaz, K. M.Nadim Shahariar, and Muhammed J.A. Patwary. Class Imbalance Problems in Machine Learning: A Review of Methods And Future Challenges. ACM Int. Conf. Proceeding Ser., pages 485–490, 2022. doi:10.1145/3542954.3543024

    2022

  11. [19]

    RGAN-EL: A GAN and ensemble learning-based hybrid approach for imbalanced data classification

    Hongwei Ding, Yu Sun, Zhenyu Wang, Nana Huang, Zhidong Shen, and Xiaohui Cui. RGAN-EL: A GAN and ensemble learning-based hybrid approach for imbalanced data classification. Inf. Process. Manag., 60(2):1–20,

  12. [20]

    Generative adversarial network in mechanical fault diagnosis under small sample: A systematic review on applications and future perspectives

    Tongyang Pan, Jinglong Chen, Tianci Zhang, Shen Liu, Shuilong He, and Haixin Lv. Generative adversarial network in mechanical fault diagnosis under small sample: A systematic review on applications and future perspectives. ISA Trans., 128:1–10, 2022. ISSN 00190578. doi:10.1016...

    2022 doi

  13. [21]

    Augmentation of Decision Tree Model Through Hyper-Parameters Tuning for Monitoring of Cutting Tool Faults Based on Vibration Signatures

    Abhishek D Patange, Sujit S Pardeshi, R Jegadeeshwaran, Ameya Zarkar, and Kshitiz Verma. Augmentation of Decision Tree Model Through Hyper-Parameters Tuning for Monitoring of Cutting Tool Faults Based on Vibration Signatures. J. Vib. Eng. Technol., 11(8):3759–3777, 2023. ISSN ...

    2023 doi

  14. [22]

    Impact of noise model on the performance of algorithms for fault diagnosis in rolling bearings

    Fabrizio Pancaldi, Luca Dibiase, and Marco Cocconcelli. Impact of noise model on the performance of algorithms for fault diagnosis in rolling bearings. Mech. Syst. Signal Process., 188(November 2022):109975, 2023. ISSN 10961216. doi:10.1016/j.ymssp.2022.109975

    2022

  15. [23]

    Iterative minority oversampling and its ensem- ble for ordinal imbalanced datasets

    Ning Wang, Zhong Liang Zhang, and Xing Gang Luo. Iterative minority oversampling and its ensem- ble for ordinal imbalanced datasets. Eng. Appl. Artif. Intell. , 127(PA):107211, 2024. ISSN 09521976. doi:10.1016/j.engappai.2023.107211

    2024

  16. [24]

    Efficient hybrid oversampling and intelligent undersampling for imbalanced big data classification

    Carla Vairetti, José Luis Assadi, and Sebastián Maldonado. Efficient hybrid oversampling and intelligent undersampling for imbalanced big data classification. Expert Syst. Appl., 246(February 2023):123149, 2024. ISSN 09574174. doi:10.1016/j.eswa.2024.123149

    2023

  17. [25]

    kNN approach to unbalanced data distributions: a case study involving information extraction

    Jianping Zhang and Inderjeet Mani. kNN approach to unbalanced data distributions: a case study involving information extraction. In Proc. Work. Learn. from imbalanced datasets , volume 126, 2003. URL https: //www.bibsonomy.org/bibtex/2cf4d2ac8bdac874b3d4841b4645a5a90/diana

    2003

  18. [26]

    Foundations of data imbalance and solutions for a data democracy

    Ajay Kulkarni, Deri Chong, and Feras A Batarseh. Foundations of data imbalance and solutions for a data democracy. Data Democr. Nexus Artif. Intell. Softw. Dev. Knowl. Eng., pages 83–106, 2020. doi:10.1016/B978-0- 12-818366-3.00005-8

    2020 doi

  19. [27]

    Chawla, Kevin W

    Nitesh V . Chawla, Kevin W. Bowyer, Lawrence O. Hall, and W. Philip Kegelmeyer. SMOTE: Synthetic Minority Over-sampling Technique. J. Artif. Intell. Res. , 16(Sept. 28):321–357, 2002. ISSN 10769757. doi:10.48550/arXiv.1106.1813

    2002 doi arXiv

  20. [28]

    Wang, Stefanka S

    Alex X. Wang, Stefanka S. Chukova, and Binh P. Nguyen. Synthetic minority oversampling using edited displacement-based k-nearest neighbors. Appl. Soft Comput. , 148(October):110895, 2023. ISSN 15684946. doi:10.1016/j.asoc.2023.110895

    2023

  21. [29]

    Enhancing and improving the performance of imbalanced class data using novel GBO and SSG: A comparative analysis

    Md Manjurul Ahsan, Md Shahin Ali, and Zahed Siddique. Enhancing and improving the performance of imbalanced class data using novel GBO and SSG: A comparative analysis. Neural Networks, 173(January): 106157, 2024. ISSN 18792782. doi:10.1016/j.neunet.2024.106157

    2024

  22. [30]

    Garcia, and Shutao Li

    Haibo He, Yang Bai, Edwardo A. Garcia, and Shutao Li. ADASYN: Adaptive synthetic sampling approach for im- balanced learning. Proc. Int. Jt. Conf. Neural Networks, (3):1322–1328, 2008. doi:10.1109/IJCNN.2008.4633969

    2008

  23. [31]

    KNNOR: An over- sampling technique for imbalanced datasets

    Ashhadul Islam, Samir Brahim Belhaouari, Atiq Ur Rehman, and Halima Bensmail. KNNOR: An over- sampling technique for imbalanced datasets. Appl. Soft Comput. , 115:108288, 2022. ISSN 15684946. doi:10.1016/j.asoc.2021.108288

    2022

  24. [32]

    Wang, Stefanka S

    Alex X. Wang, Stefanka S. Chukova, and Binh P. Nguyen. Implementation and analysis of centroid displacement- based k-nearest neighbors. In Weitong Chen, Lina Yao, Taotao Cai, Shirui Pan, Tao Shen, and Xue Li, editors, Advanced Data Mining and Applications , pages 431–443, Cham...

    2022

  25. [33]

    R-WDLS: An efficient security region oversampling technique based on data distribution

    Liyan Jia, Zhiping Wang, Pengfei Sun, and Zhaohui Xu. R-WDLS: An efficient security region oversampling technique based on data distribution. Appl. Soft Comput. , 154(February):111376, 2024. ISSN 15684946. doi:10.1016/j.asoc.2024.111376

    2024

  26. [34]

    Pereira, Anabela Afonso, and Fátima Melo Medeiros

    Dulce G. Pereira, Anabela Afonso, and Fátima Melo Medeiros. Overview of Friedmans Test and Post-hoc Analysis. Commun. Stat. Simul. Comput., 44(10):2636–2653, 2015. ISSN 15324141. doi:10.1080/03610918.2014.931971

    2015

  27. [35]

    A comprehensive comparison among metaheuristics (MHs) for geohazard modeling using machine learning: Insights from a case study of landslide displacement prediction

    Junwei Ma, Ding Xia, Yankun Wang, Xiaoxu Niu, Sheng Jiang, Zhiyang Liu, and Haixiang Guo. A comprehensive comparison among metaheuristics (MHs) for geohazard modeling using machine learning: Insights from a case study of landslide displacement prediction. Eng. Appl. Artif. Int...

    2022

  28. [36]

    Hybrid Firefly Optimised Ensemble Classification for Drift- ing Data Streams with Imbalance

    Blessa Binolin Pepsi M and Senthil Kumar N. Hybrid Firefly Optimised Ensemble Classification for Drift- ing Data Streams with Imbalance. Knowledge-Based Syst., 288(January):111500, 2024. ISSN 09507051. doi:10.1016/j.knosys.2024.111500

    2024

  29. [37]

    Mawuli, Qinli Yang, and Junming Shao

    Salah Ud Din, Aman Ullah, Cobbinah B. Mawuli, Qinli Yang, and Junming Shao. A reliable adaptive prototype- based learning for evolving data streams with limited labels. Inf. Process. Manag., 61(1):103532, 2024. ISSN 03064573. doi:10.1016/j.ipm.2023.103532

    2024

  30. [38]

    Statistical comparisons of classifiers over multiple data sets

    Janez Demšar. Statistical comparisons of classifiers over multiple data sets. J. Mach. Learn. Res., 7:1–30, 2006. ISSN 15337928. doi:10.5555/1248547.1248548

    2006

  31. [39]

    Electricity theft detection with automatic labeling and enhanced rusboost classification using differ- ential evolution and jaya algorithm

    Sana Mujeeb, Nadeem Javaid, Abrar Ahmed, Sardar Muhammad Gulfam, Umar Qasim, Muhammad Shafiq, and Jin-Ghoo Choi. Electricity theft detection with automatic labeling and enhanced rusboost classification using differ- ential evolution and jaya algorithm. IEEE Access, 9:128521–12...

    2021

  32. [40]

    Resampling imbalanced data for network intrusion detection datasets

    Sikha Bagui and Kunqi Li. Resampling imbalanced data for network intrusion detection datasets. Journal of Big Data, 8(1):6, Jan 2021. ISSN 2196-1115. doi:10.1186/s40537-020-00390-x. URL https://doi.org/10. 1186/s40537-020-00390-x

    2021 doi

  33. [41]

    Improved multi-class classification approach for imbalanced big data on spark

    Tinku Singh, Riya Khanna, Satakshi, and Manish Kumar. Improved multi-class classification approach for imbalanced big data on spark. The Journal of Supercomputing, 79(6):6583–6611, Apr 2023. ISSN 1573-0484. doi:10.1007/s11227-022-04908-3. URL https://doi.org/10.1007/s11227-022...

    2023 doi

  34. [2020]

    doi:10.1177/1550147720916404

    ISSN 15501477. doi:10.1177/1550147720916404

    doi

  35. [2023]

    doi:10.1016/j.ipm.2022.103235

    ISSN 03064573. doi:10.1016/j.ipm.2022.103235

    2022

Pith tools

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