Figure 5

Classification results are expressed as AUC-ROC curves for seven seizure attack types. (a). Results of the SVM model. (b). Results by using the CNN model. (c). Results by using the GNN model. (d). Transformer detection results. (e). Results by the GGN model. (f). Accuracy scores of two compound performance metrics for five distinct machine learning and neural network models for the classification of epileptic seizure attack types.