|
|
|
| Dynamic Motor Imagery Classification with Decision Fusion Based on Linear Discriminant Analysis |
| FU Rong-rong1,LI Peng1,LIU Chong2,ZHANG Yang3 |
1. School of Electrical Engineering, Yanshan University, Qinhuangdao, Hebei 066004, China
2. College of Mechanical Engineering and Automation, Northeastern University, Shenyang, Liaonin 110819, China
3. Design and Research Institute of Shenyang Machine Tool (Group) Co., Shenyang, Liaonin 110142, China |
|
|
|
|
Abstract The recognition and classification of electroencephalogram (EEG) signals has always been a hot topic in brain computer interface technology. Aiming at the problem that single classifier can not make good use of features and the adaptability of classifiers, it is difficult to further improve the recognition accuracy. A classification decision level fusion strategy based on linear discriminant analysis (LDA) is proposed to improve the classification accuracy of brain computer interface system. Firstly, the false test features of the two classifiers are separated, and the more likely correct decision is selected from the two methods to improve the classification accuracy; secondly, in order to measure the uncertainty of each decision, the maximum and the second largest correlation coefficient with the corresponding classifier is used to extract the feature vector. Based on this idea, a new decision selector is proposed, the method combines two LDA based algorithms to select more likely accurate decisions, so as to improve the classification accuracy of EEG signals. The experimental results show that this method can achieve better classification accuracy in motion imagery data classification by combining with the algorithm with similar accuracy.
|
|
Received: 18 February 2021
Published: 18 May 2022
|
|
|
|
|
|
［1］Li L L, Xu G H, Zhang F, et al. Relevant Feature Integration and Extraction for Single-Trial Motor Imagery Classification ［J］. Frontiers in Neuroscience, 2017 (11): 371.
［2］胡春海, 信思旭, 刘斌, 等. 基于小波变换和盲源分离的P300识别算法研究 ［J］. 计量学报, 2017, 38(2): 242-246.
Hu C H, Xin S X, Liu B, et al. Study on Recognition Algorithm of P300 Based on Wavelet Transform and Blind Source Separation ［J］. Acta metrologica Sinica, 2017, 38(2): 242-246.
［3］Lemm S, Blankertz B, Curio G, et al. Spatio-Spectral Filter for Improved Classification of Single-Trial EEG ［J］. IEEE transactions on bio-medical engineering, 2005, 52(9): 1541-1548.
［4］李晓辉, 黎湘, 郭桂蓉. 基于LDA算法的一维距离像特征提取 ［J］. 国防科技大学学报, 2005, 27(6): 72-76.
Li X H, Li X, Guo G R. One-dimensional range profile feature extraction based on LDA algorithm ［J］. Journal of National University of Defense Technology, 2005, 27(6): 72-76.
［5］Dixon S J, Brereton R G. Comparison of performance of five common classifiers represented as boundary methods: euclidean distance to centroids, linear discriminant analysis, quadratic discriminant analysis, learning vector quantization and support vector machines, as dependent on data structure ［J］. Chemometrics and Intelligent Laboratory Systems, 2009, 95(1): 1-17.
［6］Zhou Y, Zhang B X, Li G R, et al.
GD-RDA: A New Regularized Discriminant Analysis for High-Dimensional Data［J］. Journal of Computational Biology, 2017, 24(11): 0029.
［7］何群, 杜硕, 王煜文, 等. 基于变分模态分解与深度信念网络的运动想象分类识别研究 ［J］. 计量学报, 2020, 41(1): 90-99.
He Q, Du S, Wang Y W, et al. The Classification of EEG lnduced by Motor lmagery Based on Variational Mode Decomposition and Deep Belief Network ［J］. Acta metrologica Sinica, 2020, 41(1): 90-99.
［8］Liu Y T, Pal N R, Marathe A R, et al. Fuzzy Decision-Making Fuser (FDMF) for Integrating Human-Machine Autonomous (HMA) Systems with Adaptive Evidence Sources ［J］. Frontiers in Neuroscience, 2017, 11: 332.
［9］万柏坤, 刘延刚, 明东, 等. 基于脑电特征的多模式想象动作识别 ［J］. 天津大学学报, 2010, 43(10): 895-890.
Wan B K, Liu Y G, Ming D, et al. Multi-modal Imaginative Action Recognition Based on EEG Features ［J］. Journal of Tianjin University, 2010, 43(10): 895-890.
［10］ Ge S, Wang R M, Yu D. Classification of four-class motor imagery employing single-channel electroencephalography ［J］. Computer Science, Medicine, 2014, 9(6): 1-7.
［11］汪朝海, 蔡晋辉, 曾九孙. 基于经验模态分解和主成分分析的滚动轴承故障诊断研究 ［J］. 计量学报, 2019, 40(6): 1077-1082.
Wang C H, Cai J H, Zeng j S. Research on rolling bearing fault diagnosis based on empirical mode decomposition and principal component analysis ［J］. Acta metrologica Sinica, 2019, 40 (6): 1077-1082.
［12］Huang R H R, Liu Q L Q, Lu H L H, et al. Solving the small sample size problem of LDA ［C］// 16 Th International Conference on Pattern Recognition. IEEE Computer Society, 2002.
［13］Bhattacharyya S, Khasnobish A, Chatterjee S, et al. Performance analysis of LDA, QDA and KNN algorithms in left-right limb movement classification from EEG data ［C］// International Conference on Systems in Medicine & Biology. 2011.
［14］Hekmatmanesh A, Jamaloo F, Wu H, et al. Common spatial pattern combined with kernel linear discriminate and generalized radial basis function for motor imagery-based brain computer interface applications ［C］// 2017.
［15］付荣荣, 鲍甜恬, 田永胜, 等. 基于子成分分解的脑电信号去噪方法比较研究 ［J］. 计量学报, 2019, 40(4): 708-713.
Fu R R, Bao T T, Tian Y S, et al. Comparative study on EEG signal denoising methods based on sub component decomposition ［J］. Acta metrologica Sinica, 2019, 40 (4): 708-713.
［16］Mansor M N, Rejab M N. A computational model of the infant pain impressions with Gaussian and Nearest Mean Classifier. ［C］// IEEE International Conference on Control System IEEE, 2013.
［17］Loog M. Constrained parameter estimation for semi-supervised learning: the case of the nearest mean classifier. ［C］// Machine Learning & Knowledge Discovery in Databases, European Conference, Ecml Pkdd, Barcelona, Spain, SeptemberDBLP, 2011.
［18］Veenman C J, Tax D. A Weighted Nearest Mean Classifier for Sparse Subspaces. ［C］// IEEE Computer Society Conference on Computer Vision & Pattern Recognition IEEE Computer Society, 2005. |
|
|
|
2022, Vol. 43 
