Imbalanced class distribution is an inherent problem in many real-world classification tasks where the minority class is the class of interest. Many conventional statistical and machine learning classification algorithms are subject to frequency bias, and learning discriminating boundaries between the minority and majority classes could be challenging. To address the class distribution imbalance in deep learning, we propose a class rebalancing strategy based on a class-balanced dynamically weighted loss function where weights are assigned based on the class frequency and predicted probability of ground-truth class. The ability of dynamic weighting scheme to self-adapt its weights depending on the prediction scores allows the model to adjust for instances with varying levels of difficulty resulting in gradient updates driven by hard minority class samples. We further show that the proposed loss function is classification calibrated. Experiments conducted on highly imbalanced data across different applications of cyber intrusion detection (CICIDS2017 data set) and medical imaging (ISIC2019 data set) show robust generalization. Theoretical results supported by superior empirical performance provide justification for the validity of the proposed dynamically weighted balanced (DWB) loss function.

Cross‐channel maximum and minimum are used to monitor real‐time electroencephalogram signals in 14 channels. On detection of a possible change, multivariate empirical mode decomposed the last 2 s of the signal into narrow‐band intrinsic mode functions. Common spatial pattern is then utilised to create discriminating features for classification purpose. Logistic regression, artificial neural network, and support vector machine classifiers all could detect the eye state change with 83.4% accuracy in <2 s. This algorithm provides a valuable improvement in comparison with a recent procedure that took about 20 min to classify new instances with 97.3% accuracy. Application of the introduced algorithm in the real‐time eye state classification is promising. Increasing the training examples could even improve the accuracy of the classification analytics.