摘要
为了提高分类识别率,提出了一种新的公共空间算法—时变空间滤波算法。该算法中,公共空间滤波器的值不再是常数,而将随着时间而变化,它能够有效提高任务的识别率。首先给出公共空间模型的数学公式,然后进行时变公共空间滤波器的数学求解,试验结果表明时变公共空间滤波器在BCI的实用性,能够达到很好的识别效果。
In order to improve the accuracy of classified the brain state,a time-variant spatial filter algorithm was put forward.In this algorithm,the parameter of the spatial pattern filter isn＇t fixed,but varies with time.So the algorithm can improve the accuracy of classified the brain state.A short description of the CSP algorithm was proposed.Then the time-variant spatial filter algorithm was introduced with an elaborate description.Finally,the time-variant spatial filter algorithm was applied in BCI data set to prove the feasibility of the algorithm.The result illustrates the performance and validity of the algorithm.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2009年第23期7569-7571,7576,共4页
Journal of System Simulation
基金
湖南省自然科学基金(07JJ6045)
关键词
脑-机接口
公共空间模型
时变公共空间滤波器
脑电信号
brain-computer interface（BCI）
common spatial patterns（CSP）
time-variant spatial filter
Electroencephalogram
