کنفرانس مهندسی برق ایران

صفحه اصلی / سی و سومین کنفرانس بین المللی مهندسی برق

A Multi-domain Fuzzy Ensemble Approach for Epileptic Seizure Detection

نویسندگان :

Samin Shahraki¹ Alireza Hajabdollah javaheri² Mahdi Pourgholi³ Pedram Safarpour⁴

1- Shahid Beheshti University 2- Institute for Research in Fundamental Sciences (IPM), 3- Shahid Beheshti University 4- Shahid Beheshti University

کلمات کلیدی :

Epilepsy،Seizure،ANFIS،Ensemble،EEG،Fuzzy

چکیده :

Accurate diagnosis of epilepsy and its associated seizures is crucial in the treatment plan. The most common clinical test is electroencephalogram signals generated from the brain’s electrical activity. Identifying patterns in these signals provides useful information about epilepsy. This research introduces a novel framework using adaptive neuro-fuzzy inference systems (ANFIS) to recognize patterns and categorize EEG signals into seizure, non-seizure, and healthy. Feature extraction was applied by considering time, frequency, and wavelet domains, and the most relevant features in each domain were identified. Experimental results highlight that wavelet transform alone cannot capture the most relevant features. To address this limitation by incorporating multiple feature domains a meta-classifier was employed to combine ANFIS models based on subtractive clustering, forming an ensemble pipeline. The proposed method achieved an accuracy of 97.40% while providing high interpretability, enabled by the fuzzy decision-making process inherent in the ANFIS models, effectively addressing challenges such as the complex fractal and oscillatory similarities between seizure and non-seizure signals.