Table 2 Classification accuracy of five feature sets (TD, EMD-SVD, VMD-SVD, TD + EMD-SVD, TD + VMD-SVD) corresponding to four different classifiers (SVM, KNN, DT, and RF) for dataset I and Ninapro DB3.
Feature vectors | Dataset | SVM (%) | KNN (%) | DT (%) | RF (%) |
|---|---|---|---|---|---|
Time-domain | Dataset I | 81.83 ± 1.07 | 82.50 ± 0.4 | 87.40 ± 0.70 | 92.04 ± 0.37 |
Ninapro DB3 | 50.26 ± 1.25 | 66.32 ± 2.53 | 64.68 ± 1.09 | ||
EMD-SVD | Dataset I | 75.24 ± 1.12 | 83.02 ± 1.05 | 86.57 ± 0.80 | 90.10 ± 0.44 |
Ninapro DB3 | 76.71 ± 0.95 | 50.59 ± 1.13 | 61.88 ± 1.87 | ||
VMD-SVD | Dataset I | 86.85 ± 1.09 | 97.45 ± 0.33 | 93.37 ± 0.52 | 97.00 ± 0.21 |
Ninapro DB3 | 93.25 ± 0.52 | 89.81 ± 0.75 | 95.66 ± 1.30 | ||
Time-domain + EMD-SVD | Dataset I | 83.28 ± 1.25 | 86.66 ± 0.62 | 89.30 ± 0.61 | 92.47 ± 1.36 |
Ninapro DB3 | 80.41 ± 1.9 | 88.96 ± 0.79 | 86.65 ± 1.33 | ||
Time-domain + VMD-SVD | Dataset I | 90.35 ± 1.03 | 91.73 ± 0.66 | 92.33 ± 0.46 | 92.41 ± 0.46 |
Ninapro DB3 | 69.73 ± 0.80 | 91.38 ± 1.53 | 93.75 ± 0.85 |
