Table 2 Ablation study results for enhanced LSTM model.
From: Advancements in noise reduction for wheel speed sensing using enhanced LSTM models
Parameter | Setting | MSE | SNR (dB) |
|---|---|---|---|
Optimizer | Adam (Baseline) | 0.000662 | 31.797619 |
SGD | 0.001500 | 30.219953 | |
RMSprop | 0.000902 | 31.042187 | |
Max training epochs | 30 | 0.000785 | 31.248965 |
50 (Baseline) | 0.000662 | 31.797619 | |
100 | 0.000711 | 31.509342 | |
Mini-batch size | 16 | 0.000900 | 31.003421 |
32 (Baseline) | 0.000662 | 31.797619 | |
64 | 0.000785 | 31.389022 | |
Initial learning rate | 0.0005 | 0.000935 | 30.765432 |
0.001 (Baseline) | 0.000662 | 31.797619 | |
0.01 | 0.001110 | 29.847531 | |
Data shuffling | None | 0.000780 | 31.217954 |
Every-Epoch (Baseline) | 0.000662 | 31.797619 |
