Application of SVM and wavelet analysis in EEG classification.
- Author:
Jianlin ZHAO
1
;
Weidong ZHOU
;
Kai LIU
;
Dongmei CAI
Author Information
1. School of Information Science and Engineering, Shandong University, Jinan 250100, China.
- Publication Type:Journal Article
- MeSH:
Electroencephalography;
classification;
methods;
Epilepsy;
diagnosis;
physiopathology;
Humans;
Support Vector Machine;
Wavelet Analysis
- From:
Journal of Biomedical Engineering
2011;28(2):277-279
- CountryChina
- Language:Chinese
-
Abstract:
We employed two methods of support vector machines (SVM) combined with two kinds of wavelet analysis to classify these EEG signals, on the basis of the different profiles, energy, and frequency characteristics of the EEG during the seizures. One method was to classify these signals using waveform characteristics of the EEG signal. The other was to classify these signals based on fluctuation index and variation coefficient of the EEG signal. We compared the classification accuracies of these two methods with the intermittent EEG and epileptic EEG. The results of the experiments showed that both the two methods for distinguishing epileptic EEG and interictal EEG can achieve an effective performance. It was also confirmed that the latter, the method based on the fluctuation index and variation coefficient, possesses a better effect of classification.
