TY - JOUR
T1 - Classification of multi-class motor imagery with a novel hierarchical SVM algorithm for brain–computer interfaces
AU - Dong, Enzeng
AU - Li, Changhai
AU - Li, Liting
AU - Du, Shengzhi
AU - Belkacem, Abdelkader Nasreddine
AU - Chen, Chao
N1 - Funding Information:
This work was financially supported by the National Natural Science Foundation of China (61502340, 61172185), Natural Science Foundation of Tianjin City (15JCYBJC51800), and Higher School Science and Technology Development Fund Planning Project of Tianjin City (20120829).
Publisher Copyright:
© 2017, International Federation for Medical and Biological Engineering.
PY - 2017/10/1
Y1 - 2017/10/1
N2 - Pattern classification algorithm is the crucial step in developing brain–computer interface (BCI) applications. In this paper, a hierarchical support vector machine (HSVM) algorithm is proposed to address an EEG-based four-class motor imagery classification task. Wavelet packet transform is employed to decompose raw EEG signals. Thereafter, EEG signals with effective frequency sub-bands are grouped and reconstructed. EEG feature vectors are extracted from the reconstructed EEG signals with one versus the rest common spatial patterns (OVR-CSP) and one versus one common spatial patterns (OVO-CSP). Then, a two-layer HSVM algorithm is designed for the classification of these EEG feature vectors, where “OVO” classifiers are used in the first layer and “OVR” in the second layer. A public dataset (BCI Competition IV-II-a)is employed to validate the proposed method. Fivefold cross-validation results demonstrate that the average accuracy of classification in the first layer and the second layer is 67.5 ± 17.7% and 60.3 ± 14.7%, respectively. The average accuracy of the classification is 64.4 ± 16.7% overall. These results show that the proposed method is effective for four-class motor imagery classification.
AB - Pattern classification algorithm is the crucial step in developing brain–computer interface (BCI) applications. In this paper, a hierarchical support vector machine (HSVM) algorithm is proposed to address an EEG-based four-class motor imagery classification task. Wavelet packet transform is employed to decompose raw EEG signals. Thereafter, EEG signals with effective frequency sub-bands are grouped and reconstructed. EEG feature vectors are extracted from the reconstructed EEG signals with one versus the rest common spatial patterns (OVR-CSP) and one versus one common spatial patterns (OVO-CSP). Then, a two-layer HSVM algorithm is designed for the classification of these EEG feature vectors, where “OVO” classifiers are used in the first layer and “OVR” in the second layer. A public dataset (BCI Competition IV-II-a)is employed to validate the proposed method. Fivefold cross-validation results demonstrate that the average accuracy of classification in the first layer and the second layer is 67.5 ± 17.7% and 60.3 ± 14.7%, respectively. The average accuracy of the classification is 64.4 ± 16.7% overall. These results show that the proposed method is effective for four-class motor imagery classification.
KW - Common spatial pattern
KW - Electroencephalography (EEG)
KW - Hierarchical support vector machine (HSVM)
KW - Motor imagery
UR - http://www.scopus.com/inward/record.url?scp=85013770887&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85013770887&partnerID=8YFLogxK
U2 - 10.1007/s11517-017-1611-4
DO - 10.1007/s11517-017-1611-4
M3 - Article
AN - SCOPUS:85013770887
SN - 0140-0118
VL - 55
SP - 1809
EP - 1818
JO - Medical and Biological Engineering and Computing
JF - Medical and Biological Engineering and Computing
IS - 10
ER -