Journal of Electrical Systems and Information Technology, Год журнала: 2025, Номер 12(1)

Опубликована: Май 21, 2025

Abstract A novel hybrid approach combining neighborhood component analysis (NCA) and metaheuristic optimization algorithms is proposed to improve the classification accuracy of electromyography (EMG) signals while reducing feature set size computational time. EMG were collected from six neck shoulder muscles, a total 23 features extracted, including 17 time domain 6 frequency features. The extracted underwent pre-processing selection using both filter-based wrapper-based techniques. Among evaluated algorithms, gray wolf (GWO) equilibrium (EO) showed best performance. approaches, NCA-GWO NCA-EO, combined strengths NCA for ranking with search efficiency GWO EO. These methods achieved high accuracy, reduced subsets, improved compared individual methods. Statistical validation Wilcoxon signed-rank tests confirmed that performance NCA-EO was statistically comparable. Evaluation metrics such as mean squared error (MSE), kappa coefficient, Fisher’s score (F-score) demonstrated robustness reliability results suggest have significant potential real-time applications.

Язык: Английский