| # 🏆 BEST PERFORMANCE MODEL (Lowest MAE) | |
| python train_lstm.py --csv test1.csv --epochs 20 --batch_size 128 --hidden_size 256 --bidirectional --loss_type weighted_huber --model_out lstm_model_v3.pt | |
| ================================================== | |
| EVALUATION METRICS | |
| ================================================== | |
| MAE (Mean Absolute Error): 4.1815 mph | |
| RMSE (Root Mean Square Error): 7.6657 mph | |
| MAPE (Mean Absolute Percentage Error): 8.20% | |
| R² (Coefficient of Determination): 0.4294 | |
| Speed Range Performance: | |
| Low ≤30: 11.5730 mph (n=1889) | |
| Medium 30-60: 7.1722 mph (n=15717) | |
| High ≥60: 3.2165 mph (n=63171) | |
| Saving predictions to test_predictions.csv | |
| Predictions saved with 80539 rows |