EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Nonlinear Chaotic Characterization of EEG Signals for Automated Epileptic Seizure Detection

Maha A. Abdul-Jabar, Dhafar Fakhry Hasan, Omar I. Dallal Bashi

International Journal on Advanced Science, Engineering and Information Technology · 28.04.2026 · 16 (2) · 611-618

DOI: 10.18517/ijaseit.16.2.13792 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal on Advanced Science, Engineering and Information Technology, online erschienen am 28.04.2026. Verfasst von Maha A. Abdul-Jabar und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Epilepsy is a chronic neurological disorder which is associated with intermittent and recurrent seizures due to epileptic discharges in the brain. A long-term electroencephalogram (EEG) with patient observation is the conventional method used to diagnose epilepsy, which can be a time-consuming, subjective, and error-prone process. In this work, a chaotic-machine learning hybrid model is proposed to automatically diagnose epileptic seizures by combining nonlinear dynamics analysis with sequence modeling. The proposed system can learn chaotic features such as the Hurst Exponent, Largest Lyapunov Exponent, Recurrence Quantification Analysis, and Fractal Dimension, which capture the non-linear, chaotic nature of the EEG signal. These features are used to train a Support Vector Machine (SVM) and a Long Short-Term Memory (LSTM) to improve the classification of normal and epileptic brain states. The proposed system was validated experimentally on a publicly available EEG database, comprising the EEG recordings from five categories of healthy and epileptic patients, and attained a mean accuracy of 98.9% using the SVM. The LSTM model led to an improvement in the overall classification, with an accuracy of 99.3%, a recall of 99.2%, a precision of 99.1%, and an F1-score of 99.1%. The results demonstrate that the interpretability and stability of chaotic EEG analysis can be improved by incorporating LSTM and sequence learning. This combination, therefore, provides a strong platform for developing real-time, patient-free, and practical tools for epilepsy diagnosis and monitoring for smart healthcare and wearable neurodiagnostic systems.

Zitieren

APA

Abdul-Jabar, M. A., Hasan, D. F., & Bashi, O. I. D. (2026). Nonlinear Chaotic Characterization of EEG Signals for Automated Epileptic Seizure Detection. International Journal on Advanced Science, Engineering and Information Technology, 16(2), 611-618. https://doi.org/10.18517/ijaseit.16.2.13792

Vancouver

Abdul-Jabar MA, Hasan DF, Bashi OID. Nonlinear Chaotic Characterization of EEG Signals for Automated Epileptic Seizure Detection. International Journal on Advanced Science, Engineering and Information Technology. 2026;16(2):611-618. doi:10.18517/ijaseit.16.2.13792

BibTeX
@article{AbdulJabar2026,
  author = {Abdul-Jabar, Maha A. and Hasan, Dhafar Fakhry and Bashi, Omar I. Dallal},
  title = {Nonlinear Chaotic Characterization of EEG Signals for Automated Epileptic Seizure Detection},
  journal = {International Journal on Advanced Science, Engineering and Information Technology},
  year = {2026},
  volume = {16},
  number = {2},
  pages = {611--618},
  doi = {10.18517/ijaseit.16.2.13792}
}
Kostenlos starten →