EEG-Based Seizure Prediction Via GhostNet and Imbalanced Learning

Seizure prediction task based on electroencephalogram (EEG) is a typical imbalanced classification problem since the preictal samples are much less than the interictal samples. Resampling is a common method to solve class-imbalance problems in the area of seizure prediction. However, up-sampling has the deficiency of overfitting minority class, and down-sampling tends to discard important information. In addition, many seizure prediction algorithms depend on artificially designed features, which is a time-consuming and laborious process. To this end, we propose an end-to-end ghost network (GhostNet) and class rebalanced loss (CRB-loss) function. First, the GhostNet with 1-D convolution mainly uses linear operations instead of traditional convolution to generate feature maps, which improves model performance and reduces the amount of parameters. In addition, 1-D convolution is adopted to better reduce the data dimension. Then, cost-sensitive learning is utilized to solve the problem of data imbalance. Specifically, a simple but effective CRB-loss function is proposed to increase the misclassification cost of preictal data, which makes classifier to focus more on minority samples and classifies them correctly. Experimental performance indicates the proposed approach can achieve remarkable performance gains in seizure prediction task.