Research Article
Feature Extraction of Surface Electromyography Using Wavelet Weighted Permutation Entropy for Hand Movement Recognition
Table 3
Summary of average recognition accuracies (×100%).
| Method | Features | SVM | BPNN |
| Without decomposition | PE | 0.86 | 0.75 | WPE | 0.90 | 0.82 |
| With wavelet decomposition | Wavelet PE feature set (all sub-bands) | 1.00 | 0.97 | WWPE feature set (all sub-bands) | 1.00 | 0.98 | Wavelet PE feature set (single d3 sub-band) | 0.57 | 0.58 | WWPE feature set (single d3 sub-band) | 0.75 | 0.67 |
|
|