Performance Evaluation of Epilepsy Prediction Model Based on Seizure’ Type and Patients' Characteristics

In this paper, for the first time, we interpret the epileptic prediction model results based on seizure types, patient's gender and age, and seizure onset zone. Analyzing results indicate that seizure type, gender, age, and seizure onset zone can affect the model performance. Our findings indicate that the epileptic prediction model obtained worse performance for the patients in the range of 60 to 80 years old, male, patients with the seizure onset zone of temporal compared to younger patients, female patients, and patients with the seizure onset zone of frontal and partial lobs, respectively. Many problems may be resolved by taking these factors into account when developing an epileptic seizure prediction and detection model.