Project details: The present online application aims to utilize clinical information of patients with epilepsy (PWE) to differentiate focal epilepsy from idiopathic generalized epilepsy (IGE) by application of machine learning methods. Nine easily obtainable clinical features (based on a detailed history and physical examination) are utilised as the inputs. The classification framework benefits from multiple classifiers and their best results are exploited by a Stacking classifier to perform the final classification. The training procedure is carried out on a large database of PWE built over 14 years at the epilepsy center at Shiraz University of Medical Sciences, Iran, from 2008 until 2022. More technical details can be found in the related publication.
Input parameters: including age at seizure onset, sex, a history of febrile convulsion, a family history of epilepsy, a history of severe head injury, a history of medical comorbidity, aura with seizures, ictal-related tongue biting, and abnormal physical examination.
Aura types: 1 = No aura, 2 = Indescribable feeling, 3 = Dizziness, 4 = Fear / Nervousness / Anxiety / Adrenaline rush, 5 = Cognitive / Deja vu / Jamais vu / Forced thinking, 6 = Epigastric / Abdominal / Nausea, 7 = Elementary visual, 8 = Complex visual, 9 = Elementary auditory, 10 = Complex auditory, 11 = Olfactory, 12 = Gustatory / Taste, 13 = Left focal sensory, 14 = Right focal sensory, 15 = Other sensory, 16 = Headache, 17 = Other.
Acknowledgment: The project is supported by Shiraz University of Medical Sciences. All rights reserved to the authors. The project is conducted during the post doctoral program of Dr. Davood Fattahi, under supervision of Dr. Ali Akbar Asadi-Pooya. With deep appreciation for the efforts and supports of the other team members: Dr. Nahid Abolpour, Dr. Reza Boostani, Dr. Mohsen Farazdaghi, Dr. Mehrdad Sharifi.