Unfortunately, in previous experiments, training data and test data from electroencephalogram signals are often derived from the same cases, which may affect the clinical applicability of the classifiers.
Zhen Zhang and colleagues from Zhongshan School of Medicine, Sun Yat-sen University combined a nonlinear dynamics index -approximate entropy with a support vector machine that has strong generalization ability to classify electroencephalogram signals at epileptic interictal and ictal periods. The researchers also verified whether approximate entropy waves can be effectively applied to the automatic real-time detection of epilepsy in the electroencephalogram, and to explore its generalization ability as a classifier trained using a nonlinear dynamics index.
Their findings indicate that a nonlinear dynamics index trained classifier can effectively identify epileptic electroencephalogram signals, and has good generalization ability. This combined system is simple and fast running, which has a certain significance for the development of clinical real-time systems for detection and identification of epilepsy, and creation of a new diagnosis and treatment system of epilepsy. These results are published in the Neural Regeneration Research (Vol. 8, No. 20, 2013).
Article: ” Approximate entropy and support vector machines for electroencephalogram signal classification,” by Zhen Zhang, Yi Zhou, Ziyi Chen, Xianghua Tian, Shouhong Du, Ruimei Huang.
Neural Regen Res. 2013;8(20):1844-1852.