Design and Analysis of Explainable AI-Driven Epileptic Seizure Detection Using Machine Learning Models on the Bonn EEG Dataset

DOI

10.2991/978-94-6239-674-6_13How to use a DOI?

Keywords

Epileptic seizure detection; EEG classification; Bonn EEG dataset; machine learning; gradient boosting; random forest; feature extraction; frequency-domain analysis; nonlinear features; SMOTE balancing; explainable AI; SHAP; LIME; model interpretability; biomedical signal processing

Abstract

The problem of epileptic seizure detection is still a significant issue in clinical neuroscience because the impact of the seizure episode is unpredictable, and the underlying neural structure is complicated. The conventional diagnostic models are very reliant on the experience of the specialists in interpreting the electroencephalogram (EEG) signals, which is a time consuming and subject to human error. This paper will suggest an explainable AI-based system of automated epileptic seizures detection on the basis of the Bonn EEG dataset, focusing on the predictive performance of the machine learning models and their interpretability. The methodology involves a powerful preprocessing phase includes denoising and segmentation then statistical, temporal, and frequency-domain data are extracted. Different machine learning classifiers like the Random Forest, Support Vector Machine, Gradient Boosting and the Logistic Regression are trained and optimized to differentiate between healthy, interictal and ictal EEG samples. To achieve transparency in model decisions explainable AI methods such as SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-agnostic Explanations) are combined to determine the most significant EEG characteristics that affect the classification results. Experimental findings show that the ensemble-based models can provide high accuracy, sensitivity and specificity whereas the XAI visualizations can give clinically significant information about the feature importance like the dominance frequency bands, signal variance and entropy measurements related to seizure activity. The results indicate the twofold advantage of integrating machine learning functionality and interpretability, which is aimed at building reliable clinical decision-support systems. This study helps to enhance clear and credible AI procedures of diagnosing neurological disorders and supports the possibilities of XAI-based EEG analysis in practice.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Mukesh Kumar Bhardwaj
AU  - Avnish Shukla
AU  - Radhika Sharma
AU  - Kumud Dixit
AU  - Methily Johri
PY  - 2026
DA  - 2026/05/28
TI  - Design and Analysis of Explainable AI-Driven Epileptic Seizure Detection Using Machine Learning Models on the Bonn EEG Dataset
BT  - Proceedings of the International Conference on Sustainable Computing and Artificial Intelligence (ICSCAI 2025)
PB  - Atlantis Press
SP  - 143
EP  - 156
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6239-674-6_13
DO  - 10.2991/978-94-6239-674-6_13
ID  - Bhardwaj2026
ER  -