Epileptic Seizure Detection using EEG Signals

This repositories contains a series of notebooks for the application of machine learning classifiers on the Temple University Hospital (TUH) EEG dataset.

The dataset is publicly available here.

Libraries used

• MNE
• MNE-Features
• Numpy
• Matplotlib
• Pandas
• scikit-learn
• ast

Result

Sensitivity (or Recall), Specificity, Precision, Accuracy, and F₁ Score reported on six machine learning classifiers

Models	Sensitivity (or Recall)	Specificity	Precision	Accuracy	F1 Score
Logistic Regression	93.39	91.16	93.30	92.43	0.9300
K-Nearest Neighbour	93.05	88.85	92.00	91.28	0.9250
Decision Tree	92.06	90.04	92.52	91.20	0.9250
Random Forest	96.40	81.15	87.65	90.01	0.9183
Support Vector Machine	93.64	91.37	93.67	92.70	0.9400
Linear Discriminant Analysis	90.08	87.77	90.65	89.08	0.9050

Citation

If you find this work useful, please cite

@INPROCEEDINGS{9756061,  
    author={Khan, Irfan Mabood and Khan, Mohd Maaz and Farooq, Omar},  
    booktitle={2022 5th International Conference on Computing and Informatics (ICCI)},   
    title={Epileptic Seizure Detection using EEG Signals},   
    year={2022},  
    pages={111-117},  
    doi={10.1109/ICCI54321.2022.9756061}}
}