Naive Bayes方法识别两类运动想象脑电信号的特点考察
首发时间:2018-07-16
摘要:本文通过使用Naive Bayes方法对两类运动想象脑电信号进行分类识别,以及与SVM方法的对比,考察了Naive Bayes方法识别生理信号的能力及特点。经考察得出,Naive Bayes方法可较好地应用于生理信号的识别,且能够发挥其良好的分类性能。本文的研究从多个角度考察了Naive Bayes方法在脑电识别中的应用特点,显示出Naive Bayes方法在两类运动想象脑电识别中显著的分类作用,得到较为稳定的识别结果。研究得出该方法能够较好地结合多通道特征,较适合识别经频率分解的脑电信号。研究也展示出不同识别方法产生迥异的识别结果,也突显了方法选择的重要。
For information in English, please click here
Characteristic Investigation on Two-type Motor Imagery EEG Identification of Naive Bayes Method
Abstract:The paper investigated the ability and the characteristics of Naive Bayes method to identify the physiological signals through classifying and identifying the two-type motor imagery EEG by Naive Bayes method and through comparing with SVM method. It was concluded that the Naive Bayes method can be applied to physiological signals identification well and can give full play to its good classifying performance. The paper investigated the applying characteristics of Naive Bayes method in EEG identification from multiple aspects, showed the remarkable classifying function of this method in two-type motor imagery EEG identification, and stable results can be obtained. Result showed that Naive Bayes method can combine the multi-channel characteristics well and was suitable for the identification to the EEG signals decomposed on frequency. The research also showed wide differences of identification results produced by different identification methods, and then highlighted the importance of method selection.
Keywords: BEI (Biotic Electric Interface) EEG (electroencephalogram) Bayes SVM (support vector machine) wavelet transform
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
Naive Bayes方法识别两类运动想象脑电信号的特点考察
评论
全部评论0/1000