COMPARISON BETWEEN PCA-SVM WITH PSO-SVM FOR RECOGNIZE EMG SIGNAL Daniel S Pamungkas and Sumatri K Risandriya
Politeknik Negeri Batam
Abstract
The comparison between Support Vector Machine (SVM) combined with Principal Component Analysis (PCA) with a combination of SVM with Particle Swarm Optimization (PSO) related to signal electromyographic signals (EMG) is given in this article. A Myo sensor bracelet was used for this experiment. This device is placed in the user^s upper arm. The experiment aims to recognize the five gestures of the user^s fingers. Both algorithms use the same features in the time domain. The experiments show that the PSO-SVM algorithm is more effective in identifying the user^s finger^s gesture than the SVM one. This combination reaches about 3% higher than the PCA-SVM.
