Table 9.
Forecast 1 and forecast 2 to predict Motor ON/OFF.
| Validation | Forecast 1 − motor | Forecast 2 – motor |
|---|---|---|
| Train Accuracy (%) | 95.6 | 95.1 |
| Test Accuracy (%) | 94.74 | 94.84 |
| Batch size | 32 | 32 |
| Optimizer | Adam | Adam |
| Activation | Relu | Relu |
| Epochs | 30 | 30 |
| Learning rate | 0.0005 | 0.0005 |
| Dense layer at first level (neurons) | 20 | 20 |
| Dense layer at second level (neurons) | 10 | 10 |
| Input layer forecast 1 − motor Input layer forecast 2 − motor |
4 (curr1, temp2, hum1, act1) | 3 (temp2, hum1, act1) |
| Output layer (ON/OFF) | 2 (ON/OFF) | 2 (ON/OFF) |
| Inferencing time (ms) | 1 | 1 |
| Peak RAM Usage (bytes) | 1.4 K | 1.4 K |
| Flash Usage (bytes) | 14.9 K | 14.8 K |