摘要
如何从脑电信号中快速准确地识别出P300成分是脑-机接口研究中的一个热点问题。针对P300的识别问题,我们提出了一种将F-score特征选择与支持向量机相结合的判别方法,该方法采用F-score特征选择减少输入特征的维数,以克服支持向量机算法判别速度慢的缺点;然后借助支持向量机算法良好的分类性能实现P300的识别。本文在BCI Competition 2003的P300实验数据集上对该方法进行了验证,结果表明,在5次重复实验中该方法的识别准确率达到了100%,且判别速度与未经特征选择的传统支持向量机算法相比提高了近2倍。
How to detect the P300 component in EEG accurately and instandy is a hot problem in the research field of Brain-Computer Interface. In this paper, an algorithm based on F-score feature selection and support vector machines was introduced for P300 detection. Using F-score feature selection method, we reduced input features to overcome the shortooming of support vector machines in terms of low detection speed, and then implemented the detection of P300 oomponent with support vector machines, which have good classification performance. The algorithm was tested with a P300 dataset from the BCI competition 2003. The results showed that the algorithm achieved an accuracy of 100 % in P300 detection within five repetitions, and the detection speed of this algorithm was 2 times higher than that of the traditional support vector machines algorithm without F-score feature selection.
出处
《生物医学工程学杂志》
EI
CAS
CSCD
北大核心
2008年第1期23-26,52,共5页
Journal of Biomedical Engineering
基金
山东省自然科学基金资助项目(Y2005G12)