Bagging RCSP脑电特征提取算法

正则化共空间模式（Regularized common spatial pattern,RCSP）解决了共空间模式（Common spatial pattern,CSP）对噪声敏感的问题,但它在小样本脑电数据集中的表现并不理想.针对上述问题,本文提出了Bagging RCSP（BRCSP）算法,通过Bagging方法重复选取训练数据来构造一个个包,并提取RCSP特征,再利用线性判别分析（Linear discriminant analysis,LDA）将特征向量映射到低维空间中,最后采用最近邻（Nearest neighborhood classifier,NNC）算法判定分类结果.线下实验证...

Full description

Saved in:
Bibliographic Details
Published in自动化学报 Vol. 43; no. 11; pp. 2044 - 2050
Main Author 张毅;尹春林;蔡军;罗久飞
Format Journal Article
LanguageChinese
Published 重庆邮电大学先进制造工程学院 重庆 400065%重庆邮电大学 自动化学院 重庆 400065 2017
Subjects
Bagging
RCSP
RCSP-A
特征提取
脑电信号
特征提取
Bagging RCSP
feature extraction
RCSP-A
脑电信号
Electro encephalon graph (EEG)
Online AccessGet full text

Cover

More Information
Summary:正则化共空间模式（Regularized common spatial pattern,RCSP）解决了共空间模式（Common spatial pattern,CSP）对噪声敏感的问题,但它在小样本脑电数据集中的表现并不理想.针对上述问题,本文提出了Bagging RCSP（BRCSP）算法,通过Bagging方法重复选取训练数据来构造一个个包,并提取RCSP特征,再利用线性判别分析（Linear discriminant analysis,LDA）将特征向量映射到低维空间中,最后采用最近邻（Nearest neighborhood classifier,NNC）算法判定分类结果.线下实验证明,相比较聚合正则化共空间模式（RCSP with aggregation,RCSP-A）,BRCSP的平均准确率提高了2.92%,且方差更小,鲁棒性更好.最后,在智能轮椅平台上,10位受试者利用BRCSP算法实现左右手运动想象脑电信号控制轮椅完成＂8＂字形路径的实验,证明了该算法在脑电信号特征提取中的有效性.
Bibliography:The regularized common spatial pattern （RCSP） has solved the problem that the common spatial pattern （CSP） is sensitive to noise. However, its performance on small sample of electro encephalon graph （EEG） data set is not ideal. To deal with this problem, a Bagging RCSP （BRCSP） algorithm is proposed, which divides training samples into packets and extracts RCSP features by Bagging to choose training packets, l~lrthermore, the feature vector is projected into the lower space with linear discriminant analysis （LDA） and a classification algorithm based on nearest neighborhood classifier （NNC） is adopted. Compared to RCSP with aggregation （RCSP-A）, the accuracy of BRCSP increases by 2.92 % in average and the variance is smaller and has better robustness. Results of the experiment, in which 10 subjects control an intelligent wheelchair of a fixed ＂8＂ glyph trajectory, demonstrate that the BRCSP is effective in the EEG feature extraction.
ZHANG Yi1 YIN Chun-Lin2 CAI Jun2 LUO Jiu-Fei1
Electro encephalon graph （EEG）, fe
ISSN:0254-4156
1874-1029
DOI:10.16383/j.aas.2017.c160094