基于脑电模糊分析的睡眠分期方法研究

发布时间：2018-04-06 04:20

本文选题：睡眠分期　切入点：模糊特征　出处：《计算机工程与应用》2017年23期

【摘要】：利用脑电信号模糊特征分类的方法对睡眠进行分期研究。首先对脑电信号进行预处理,滤除干扰噪声后使用模糊熵算法、多尺度熵算法以及复杂度算法对脑电信号进行特征参数提取,采用最小二乘支持向量机(the Least Squares Support Vector Machine,LS-SVM)对特征参数进行分类,并将睡眠过程分为清醒期、浅睡期、深睡期和快速眼动期(Rapid Eye Movement,REM),获得分期正确率。最后通过上述方法对2 000组睡眠脑电样本进行睡眠分期测试,与专家人工分期结果进行比对,将复杂度输入到最小二乘支持向量机进行分类的平均正确率是92.65%,高于模糊熵和多尺度熵作为最小二乘向量机的输入时的准确率。基于模糊特征的复杂度提取的特征参数可以作为睡眠分期的有效依据,在保证准确度的前提下,降低人工成本。
[Abstract]:The stage of sleep was studied by using the fuzzy feature classification of EEG signals.Firstly, the EEG signal is preprocessed, and the fuzzy entropy algorithm, the multi-scale entropy algorithm and the complexity algorithm are used to extract the characteristic parameters of the EEG signal after filtering out the interference noise.The Least Squares Support Vector machine (LS-SVM) was used to classify the characteristic parameters, and the sleep process was divided into waking, shallow, deep and rapid Eye motion stages.Finally, 2000 groups of sleep EEG samples were tested by the above methods, and the results were compared with the results of expert artificial staging.The average accuracy of input complexity into least squares support vector machine is 92.65, which is higher than that of fuzzy entropy and multi-scale entropy as input of least squares vector machine.The feature parameters extracted based on the complexity of fuzzy features can be used as an effective basis for sleep stages and reduce the labor cost under the premise of ensuring accuracy.
【作者单位】：吉林大学仪器科学与电气工程学院;
【基金】：国家“十二五”科技支撑计划课题(No.2015BAI02B04) 吉林市科技计划项目(No.2015313013)
【分类号】：R740;TN911.7

【相似文献】