Resampling algorithms based on sample concatenation for imbalance learning

Shi, Hongbo; Zhang, Ying; Chen, Yuwen; Ji, Suqin; Dong, Yuanxiang

doi:10.1016/j.knosys.2022.108592

Cited by 19 publications

(11 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unlike other clustering-based over-sampling methods, the proposed approach applies modified density peaks clustering rather than traditional k-means clustering techniques to cluster the minority instances due to its capability of accurately identifying sub-clusters with different sizes and densities, which is beneficial for the proposed method to simultaneously accommodate for between-class and within-class imbalance issues caused by various reasons. Shi Hongbo et al [ 15 ].proposed the Re-SC model. Re-SC transforms an imbalanced training dataset in the original sample space into a concatenated dataset in a new sample space.…”

Section: Related Workmentioning

confidence: 99%

Dual generative adversarial networks based on regression and neighbor characteristics

Jia,

Lu,

Shen

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

Imbalanced data is a problem in that the number of samples in different categories or target value ranges varies greatly. Data imbalance imposes excellent challenges to machine learning and pattern recognition. The performance of machine learning models leans to be partially towards the majority of samples in the imbalanced dataset, which will further affect the effect of the model. The imbalanced data problem includes an imbalanced categorical problem and an imbalanced regression problem. Many studies have been developed to address the issue of imbalanced classification data. Nevertheless, the imbalanced regression problem has not been well-researched. In order to solve the problem of unbalanced regression data, we define an RNGRU model that can simultaneously learn the regression characteristics and neighbor characteristics of regression samples. To obtain the most comprehensive sample information of regression samples, the model uses the idea of confrontation to determine the proportion between the regression characteristics and neighbor characteristics of the original samples. According to the regression characteristics of the regression samples, an index ccr (correlation change rate) is proposed to evaluate the similarity between the generated samples and the original samples. And on this basis, an RNGAN model is proposed to reduce the similarity between the generated samples and the original samples by using the idea of confrontation.

show abstract

Section: Related Workmentioning

confidence: 99%

Dual generative adversarial networks based on regression and neighbor characteristics

Jia,

Lu,

Shen

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Some classification methods, such as Support Vector Machines (SVMs), apply kernel functions 1 to solve the linearity issue, yet, the lack of the data persists and penalizes classifiers' performances. Other studies, such as Shi et al (2022), used resampling algorithms based on sample concatenation (Re-SC). Re-SC transforms an imbalanced training dataset in the original sample space into a concatenated dataset in a new sample space.…”

Section: Class Overlapping and Within Class Imbalancementioning

confidence: 99%

“…𝑆 is inspired by the sample size formula in statistics (Cochran, 1977), in which 𝜖 is the acceptable tolerance error that can be adjusted as required, 𝑍 𝛼 is the critical value of the 𝑍 test at the significance level 𝛼 (Shi et al, 2022), 𝜙 is an indicator function, 4 and 𝑝 is the variance of a proportion denoting the percentage of a sample having a particular characteristic, 5 and 𝑘 ′ is a hyper-parameter related to the size of clusters. Here, 𝑝 denotes the proportion of objects in the given cluster  𝑖 belonging to the minority class .…”

Section: Extraction and Classification Of Representative Pointsmentioning

confidence: 99%

“…To thoroughly assess the statistical significance of AROSS's performance, we employed Welch's t-test (𝑊 𝑡𝑒𝑠𝑡 , Derrick et al (2016)). Several formal arguments substantiate the selection of 𝑊 𝑡𝑒𝑠𝑡 for benchmarking and comparing the performance of data-level resampling approaches (Zhang et al, 2008, Ellis et al, 2022, Shi et al, 2022, Darville et al, 2023. Unlike non-parametric tests such as Friedman or Wilcoxon, 𝑊 𝑡𝑒𝑠𝑡 is well-suited for handling imbalanced datasets, accommodating variations in variances among different resampling techniques.…”

Section: Statistical Analysis Using Welch's T-testmentioning

confidence: 99%

“…Unlike non-parametric tests such as Friedman or Wilcoxon, 𝑊 𝑡𝑒𝑠𝑡 is well-suited for handling imbalanced datasets, accommodating variations in variances among different resampling techniques. Its parametric nature allows it to consider both metrics average values and variances, providing a robust statistical evaluation that is particularly advantageous when dealing with realworld datasets exhibiting inherent imbalances (Shi et al, 2022). 𝑊 𝑡𝑒𝑠𝑡 , computed by the Eq.…”

Section: Statistical Analysis Using Welch's T-testmentioning

confidence: 99%

See 2 more Smart Citations