Enhancing Person Identification via EEG Channel Selection Utilizing Binary Grey Wolf Optimization Algorithm

This research tackles the task of selecting the most effective Electroencephalogram (EEG) channels for biometric identification. It treats this challenge as a binary optimization problem and employs a modified version of the Grey Wolf Optimizer (BGWO) to address this complex optimization problem. The biometric identification system using EEG integrates a Support Vector Machine classifier with a Radial Basis Function kernel (SVM-RBF). Feature extraction involves analyzing three specific auto-regressive coefficients. To evaluate the BGWO-SVM method, a standard EEG motor imagery dataset is used, measuring performance based on four criteria: Accuracy, F-Score, Recall, and Specificity. The BGWO-SVM method achieves an impressive accuracy of ۹۴.۱۳% with a streamlined setup of ۲۳ sensors and ۵ auto-regressive coefficients. Additionally, the method strategically selects electrodes, avoiding proximity to comprehensively capture essential information across the entire head. The study highlights the exceptional performance of BGWO-SVM in terms of selected channels and competitive classification accuracies, positioning it favorably among alternative meta-heuristic algorithms in the field of EEG-based biometric identification. This research contributes to ongoing EEG signal studies, emphasizing the practicality and efficiency of the proposed methodology for real-world applications.

کلیدواژه ها:

EEG ، channels selection ، Grey Wolf Optimizer ، identification ، binary optimization.