Research on Classification of Motor Imagery EEG Signals Based on LSTM and SVM
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
To address the issue of low accuracy in traditional Support Vector Machine (SVM) methods for analyzing non-linear and non-stationary electroencephalogram (EEG) signals, which leads to imprecise feature extraction and suboptimal classification performance, this paper proposes a novel EEG signal classification method based on Local Mean Decomposition (LMD), Long Short-Term Memory (LSTM) networks, and the Eel-Grouper Optimization (EGO) algorithm. Firstly, the preprocessed EEG signals are adaptively decomposed into multiple Product Functions (PFs) and a residual component using LMD, thereby obtaining more stable local features. Subsequently, the LSTM network is employed to further process and extract temporal features from the PF components, enhancing the representational capacity of the features. Finally, the EGO algorithm is utilized to optimize the parameters of the SVM, improving the classifier's accuracy and generalization ability. Experimental results demonstrate that the proposed method achieves an average classification accuracy of 90.1918% on a standard motor imagery EEG dataset, significantly outperforming other SVM-based approaches. This study provides a robust and effective solution for the classification of motor imagery EEG signals.
Fields of Study & Subjects
EEG and Brain-Computer Interfaces
Muscle activation and electromyography studies
Emotion and Mood Recognition
Publication details
- Journal / source
- IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
- Pages
- 582–585
- Publication date
- 28 January 2026
- Language
- EN