Modeling the Detection of the Prevalence of Epileptic Seizures using Machine Learning

Esraa Bakhsh; Somaiyah Alshabeer; Nahla Aljojo; Ahmed Alamri; Ammar Almutawa; Alaa Alsaig; Iqbal Alsaleh; Aisha Blfgeh; Areej Alshutayri

doi:10.37965/jait.2026.0972

Modeling the Detection of the Prevalence of Epileptic Seizures using Machine Learning

Authors

Esraa Bakhsh College of Computer Science and Engineering, Department of Information System and Technology, University of Jeddah, Saudi Arabia
Somaiyah Alshabeer College of Computer Science and Engineering, Department of Information System and Technology, University of Jeddah, Saudi Arabia
Nahla Aljojo College of Computer Science and Engineering, Department of Information System and Technology University of Jeddah, Saudi Arabia https://orcid.org/0000-0003-2501-8533
Ahmed Alamri College of Computer Science and Engineering, Department of Information System and Technology, University of Jeddah, Saudi Arabia https://orcid.org/0000-0001-9258-3503
Ammar Almutawa College of Computer Science and Engineering, Department of Information System and Technology, University of Jeddah, Saudi Arabia https://orcid.org/0009-0006-5447-5708
Alaa Alsaig College of Computer Science and Engineering, Department of Information System and Technology, University of Jeddah, Saudi Arabia
Iqbal Alsaleh Faculty of Economic and Administration, Management Information System Department, King Abdulaziz University, Saudi Arabia
Aisha Blfgeh Department of Computer Science and Artificial Intelligence, College of Computer Science and Engineering, University of Jeddah, Jeddah, Saudi Arabia https://orcid.org/0000-0001-7622-6338
Areej Alshutayri Department of Computer Science and Artificial Intelligence, College of Computer Science and Engineering, University of Jeddah, Jeddah, Saudi Arabia https://orcid.org/0000-0001-8550-0597

DOI:

https://doi.org/10.37965/jait.2026.0972

Keywords:

Decision Trees, epileptic seizures, K-Nearest Neighbors, Logistic Regression, Naive Bayes, Random Forest, XGBoost

Abstract

The frequency of epileptic seizures lead to many dimension of time frame from its beginning to its end and can be anticipated when adequate measures are in place to track and tackle it. This study presents a model designed to detect the prevalence of epileptic seizures through the application of machine learning (ML) techniques. The experimental investigation focused on exploring the early detection of epileptic seizure through the use of a dataset sourced from the UCI Machine Learning Repository. The experimental procedure entails gathering data, which is subsequently processed for analysis. Following this, the data are divided into multiple sets designated for training, testing, and validation purposes. The training involved the application of several ML algorithms, including “Logistic Regression,” “K-Nearest Neighbors,” “Decision Trees,” “Random Forest,” “Naive Bayes,” and “XGBoost.” The findings demonstrate that Random Forest and XGBoost attained accuracies of 98.31% and 98.86%, respectively. The results underscore the enhanced capabilities of these two models in comparison to others for identifying epileptic seizure. In comparison to other studies, these findings suggest that the precision of this investigation is remarkably elevated. This guarantees the development of reliable methods for predicting epileptic seizures, subsequently improving patient care and elevating quality of life.