基于深度森林的脑电注意力识别研究

Based on deep forest attention to identify research of EEG

针对脑电信号的注意力识别精度问题,本文应用深度森林的算法进行仿真研究。首先对原始脑电信号通过小波分析进行预处理去噪,然后采用深度森林的方法进行分类识别。实验分别对6位受试者在注意和非注意两种状态下的脑电信号进行分析,结果表明,对注意力状态识别的准确率达到了95%以上,同时对通用数据库中清醒和睡眠两种状态下的脑电数据进行识别,也取得了较高的识别率,结果证明了该算法对脑电信号注意力识别的准确率是可靠的。

With the aimtosolve the attention recognition accuracy for EEG signals,the algorithm of depth forest is used to simulate the EEG signal in this paper.Firstly,the original EEG signal is preprocessed by wavelet analysis,and then the method of depth forest is used to classify and identify.Through design experiment,six subjects were collected in the attention and non-attention of the two states of the EEG signal,The experimental results showed that the accuracy of attention to the state of recognition to achieve more than 95%.At the same time,the EEG data were identified in both the awake and sleep states of the general database and the recognition rate is high.The result shows that the algorithm is reliable for the accuracy of EEG attention recognition.