Research Article
sEMG-Based Neural Network Prediction Model Selection of Gesture Fatigue and Dataset Optimization
Table 2
Comparison of MSE, RMSE, and MAE for LSTM prediction models based on sEMG and its wavelet decomposition datasets.
| Prediction models | sEMG | Wavelet | MSE | RMSE | MAE | MSE | RMSE | MAE |
| LSTM-G1 | Train | 0.0214 | 0.1133 | 0.0768 | 0.0415 | 0.1709 | 0.1427 | Test | 0.0069 | 0.0696 | 0.0491 | 0.0002 | 0.1047 | 0.1027 |
| LSTM-G4 | Train | 0.033 | 0.1842 | 0.152 | 0.0148 | 0.1246 | 0.0785 | Test | 0.0261 | 0.166 | 0.1331 | 0.0004 | 0.0313 | 0.0272 |
| LSTM-G7 | Train | 0.04 | 0.0763 | 0.06 | 0.0938 | 0.1115 | 0.195 | Test | 0.0376 | 0.0717 | 0.0569 | 6.2217 | 0.0213 | 0.1377 |
|
|