论文部分内容阅读
脑电图是人脑神经元动态活动的综合表现形式,可以用来研究癫痫的脑部病理变化。本文引入多尺度排列熵(MPE)的概念,将其应用于癫痫患者和健康人的脑电图特征提取,并将所有特征参数送入支持向量机(SVM)进行分类。实验结果表明,在区分癫痫患者和健康人的脑电图时平均分类精度达100%,癫痫发作间期和发作期的平均分类精度为99.58%。与同时输入的1~5个单尺度排列熵(PE)对比分析发现,MPE比PE更能反映癫痫脑电图多尺度上的特征,能更好、更稳定地实现癫痫预测。
EEG is a comprehensive manifestation of the dynamic activity of human brain neurons, which can be used to study the brain pathological changes of epilepsy. This paper introduces the concept of multi-scale permutation entropy (MPE), applies it to EEG feature extraction of epilepsy patients and healthy people, and sends all the feature parameters into support vector machine (SVM) for classification. The experimental results showed that the average classification accuracy was 100% in distinguishing epileptic patients from healthy subjects, and the average classification accuracy was 99.58% in epileptic seizures and seizures. Compared with 1 ~ 5 single-scale permutation entropy (PE) input simultaneously, it is found that MPE can better reflect the multi-scale characteristics of epileptic EEG than PE and can achieve better and more stable epilepsy prediction.