TY - GEN
T1 - Classification of EEG Signals Based on GA-ELM Optimization Algorithm
AU - Zhang, Weiguo
AU - Lu, Lin
AU - Belkacem, Abdelkader Nasreddine
AU - Zhang, Jiaxin
AU - Li, Penghai
AU - Liang, Jun
AU - Wang, Changming
AU - Chen, Chao
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
PY - 2023
Y1 - 2023
N2 - There are many unpredictable problems in motion visualization and observation in BCI system, such as interference from external noise and visual fatigue of subjects. These problems seriously affect the performance of the whole BCI system. To solve this problem, this paper designed the experimental paradigm of imagination and observation, and built the eeg acquisition platform by combining UNITY and MATLAB. Ten healthy subjects participated in the experiment, which was divided into two stages: in the first stage, each subject was required to perform five experiments at the same time. In the second stage, after an interval of more than one month, the eeg signals of the 10 subjects were collected again (the same experimental paradigm). In pattern recognition and Hilbert huang transform time and frequency domain characteristics of extreme learning machine recognition classification based on genetic algorithm, and using the basic method of SVM algorithm and ELM comparison between the results and draw HHT and optimization algorithm of single collection of experiment acquisition signal has a significant effect, high classification rate can reach 85.3%.
AB - There are many unpredictable problems in motion visualization and observation in BCI system, such as interference from external noise and visual fatigue of subjects. These problems seriously affect the performance of the whole BCI system. To solve this problem, this paper designed the experimental paradigm of imagination and observation, and built the eeg acquisition platform by combining UNITY and MATLAB. Ten healthy subjects participated in the experiment, which was divided into two stages: in the first stage, each subject was required to perform five experiments at the same time. In the second stage, after an interval of more than one month, the eeg signals of the 10 subjects were collected again (the same experimental paradigm). In pattern recognition and Hilbert huang transform time and frequency domain characteristics of extreme learning machine recognition classification based on genetic algorithm, and using the basic method of SVM algorithm and ELM comparison between the results and draw HHT and optimization algorithm of single collection of experiment acquisition signal has a significant effect, high classification rate can reach 85.3%.
KW - BCI
KW - ELM
KW - Hilbert huang transform
KW - Motor imagination
KW - Motor observation
UR - http://www.scopus.com/inward/record.url?scp=85144213501&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85144213501&partnerID=8YFLogxK
U2 - 10.1007/978-981-19-8222-4_1
DO - 10.1007/978-981-19-8222-4_1
M3 - Conference contribution
AN - SCOPUS:85144213501
SN - 9789811982217
T3 - Communications in Computer and Information Science
SP - 3
EP - 14
BT - Human Brain and Artificial Intelligence - Third International Workshop, HBAI 2022, Held in Conjunction with IJCAI-ECAI 2022, Revised Selected Papers
A2 - Ying, Xiaomin
PB - Springer Science and Business Media Deutschland GmbH
T2 - 3rd International Workshop on Human Brain and Artificial Intelligence, HBAI 2022, held in conjunction with 31st International Joint Conference on Artificial Intelligence and 23rd European Conference on Artificial Intelligence, IJCAI-ECAI 2022
Y2 - 23 July 2022 through 23 July 2022
ER -