SW: A weighted space division framework for imbalanced problems with label noise

Li, Min; Zhou, Hao; Liu, Qun; Wang, Guoyin

doi:10.1016/j.knosys.2022.109233

Cited by 6 publications

(2 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DBSMOTE [36], kmeans-SMOTE (means-S.) [38], and RSMOTE [35] are competitive clustering-based methods. Geometric SMOTE (G-SMOTE) [40] and SMOTE-SW (S.-SW) [41] are enhanced sampling mechanisms. The core idea of their algorithm has been presented in Section II.…”

Section: Comparative Oversampling Methodsmentioning

confidence: 99%

“…Geometric SMOTE (G-SMOTE) synthesized new samples around geometric regions of the input space as an enhancement to the current data generation mechanism [40]. The SW framework performed weighted sampling by calculating the chaos of the sample space to handle imbalanced noisy datasets [41]. In conclusion, current sampling algorithms continue to have deficiencies when coping with imbalanced, noisy data sets.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

An Adaptive and Robust Method for Oriented Oversampling With Spatial Information for Imbalanced Noisy Datasets

Deng,

2023

IEEE Access

View full text Add to dashboard Cite

Imbalanced datasets have a large negative impact on the classifiers, biasing the classification results towards the majority class. Since imbalanced data distribution is an inevitable and significant challenge in the real world, many variants of SMOTE have been proposed. However, current oversampling methods still need improvement because they rely on hyperparameter optimization, overgeneralize due to emphasizing specific synthetic regions, randomly synthesize samples or suffer from noise performance degradation. To overcome the above problems, we propose an adaptive and robust method (OOSI) for oriented oversampling with spatial information to deal with imbalanced noisy datasets. OOSI is a rare adaptive and effective oversampling method that can fill the gaps of existing methods through datasetspecific spatial partitioning and information quantization, three-stage noise suppression, and spatiallyinformed generation path improvement. Firstly, a specific and adaptive clustering space is adaptively derived through the data space division of the characteristics of datasets. Then, all minority clusters are assigned a reasonable number of synthetic samples to simultaneously address intra-and inter-class imbalances by integrating the cluster samples' intra-cluster sparsity and the multi-class density information. After differentiating and identifying the noise, oriented weights are assigned based on the multi-class information level to guide the enhancement of the generation path of the synthetic samples and prevent the generation of extra noisy and overlapping samples. Extensive experiments demonstrate that the proposed algorithm outperforms 11 prominent oversampling algorithms on 11 real-world datasets with varying noise levels.INDEX TERMS Imbalanced learning, label noise, oriented oversampling,

show abstract