Table 1.
Precision results collected for different neural network architectures.
| ID | Memory Units in LSTM Layers | Number of Hidden Layers | Number of Neurons in Hidden Layers | Number of Trainable Parameters | Accuracy Test Execution 1 |
Accuracy Test Execution 2 |
Accuracy Test Execution 3 |
Training Time |
|---|---|---|---|---|---|---|---|---|
| 1 | 50 and 100 | 2 | 100 and 90 | 91.355 | 97.86% | 98.14% | 96.76% | ~29 min |
| 2 | 50 and 100 | 2 | 100 and 60 | 87.875 | 96.13% | 96.87% | 97.45% | ~30 min |
| 3 | 50 and 100 | 3 | 100, 90, and 60 | 96.365 | 96.34% | 95.33% | 95.18% | ~31 min |
| 4 | 50 and 100 | 3 | 100, 60, and 30 | 89.255 | 96.29% | 95.49% | 96.12% | ~32 min |
| 5 | 100 and 100 | 2 | 100 and 90 | 141.755 | 97.88% | 97.61% | 97.24% | ~37 min |