Epileptic electroencephalogram signal classification using wavelet energy and random forest
Epilepsy is a severe neurological disease that affects more than 50 million people worldwide. The diagnose of epilepsy can be done by analyzing the electroencephalography (EEG) recordings. Neurologist needs to find minimum one seizure (ictal) condition in the recording through a visual inspection. H...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Conference Proceeding |
| Language: | eng |
| Subjects: | |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Epilepsy is a severe neurological disease that affects more than 50 million people worldwide. The diagnose of epilepsy can be done by analyzing the electroencephalography (EEG) recordings. Neurologist needs to find minimum one seizure (ictal) condition in the recording through a visual inspection. However, this process is time-consuming and needs a lot of practice. A computer-aided diagnose to detect the ictal condition is needed to assist the neurologist in inspecting the EEG signal recordings. This study proposes the use of Wavelet energy (WE) as the feature extraction method to extract the information of the EEG signal. The calculation of WE is done in five EEG sub-bands producing five features for each signal. The features then fed to Random Forest (RF) classifier to classify three epileptic EEG conditions using five classification problems (CPs), covering the ictal vs. non-ictal condition. The combination is able to reach 100% of accuracy to classify interictal vs. ictal conditions (CP2). While the other CPs achieves accuracy from 73% to 98%. These results indicate that the use of WE is suitable for extracting the information of the EEG signal to detect epilepsy by finding the ictal condition. |
|---|---|
| ISSN: | 0094-243X 1551-7616 |