Class Imbalance
Class imbalance, the uneven distribution of classes in a dataset, hinders the performance of machine learning models by biasing them towards the majority class. Current research focuses on mitigating this imbalance through various techniques, including data resampling (oversampling minority classes, undersampling majority classes), cost-sensitive learning (assigning different misclassification costs), and algorithmic modifications (e.g., adapting loss functions, employing novel regularization methods within models like GBDTs, and using contrastive learning). Addressing class imbalance is crucial for improving the fairness, robustness, and accuracy of machine learning models across diverse applications, from medical diagnosis and financial risk assessment to environmental monitoring and traffic sign recognition.
Papers
Enhancing Clinical Predictive Modeling through Model Complexity-Driven Class Proportion Tuning for Class Imbalanced Data: An Empirical Study on Opioid Overdose Prediction
Yinan Liu, Xinyu Dong, Weimin Lyu, Richard N. Rosenthal, Rachel Wong, Tengfei Ma, Fusheng Wang
FedNoRo: Towards Noise-Robust Federated Learning by Addressing Class Imbalance and Label Noise Heterogeneity
Nannan Wu, Li Yu, Xuefeng Jiang, Kwang-Ting Cheng, Zengqiang Yan
Learning Cautiously in Federated Learning with Noisy and Heterogeneous Clients
Chenrui Wu, Zexi Li, Fangxin Wang, Chao Wu
A review of ensemble learning and data augmentation models for class imbalanced problems: combination, implementation and evaluation
Azal Ahmad Khan, Omkar Chaudhari, Rohitash Chandra