Seizure Detection Using ML Algorithms
Date:
- A novel approach for the classification of electroencephalogram (EEG) signals for seizure detection is introduced, leveraging both feature selection and channel selection techniques.
- Channel selection was conducted via the variance method. Features were then extracted from intrinsic mode functions obtained post empirical mode decomposition of the EEG signal. Feature selection involved a one-way analysis of variance test with a preset threshold probability value.
- The classification of seizure and non-seizure signals is done on selected features using Decision Tree and k-nearest neighbor algorithms.
- The proposed method achieves an impressive accuracy of 95.6\% in classifying EEG signals for seizure detection, outperforming existing methods.