Table 1 Performance comparison between LTCOP-LSTM model and other traffic flow prediction models.
Dataset | Model | Processing time (s) | Resource utilization (%) | Hyperparameters | Training time (h) | Regression loss | Convergence speed (iterations) |
|---|---|---|---|---|---|---|---|
METR-LA | LTCOP-LSTM | 25.34 | 78.4 | 8 | 15.5 | 0.0237 | 120 |
ARIMA | 30.87 | 65.2 | 5 | 10.1 | 0.0453 | 150 | |
RF | 32.45 | 62.8 | 7 | 12.3 | 0.0379 | 140 | |
PeMS | LTCOP-LSTM | 28.67 | 80.1 | 9 | 14.2 | 0.0195 | 110 |
ARIMA | 35.29 | 70.5 | 6 | 11.9 | 0.0491 | 155 | |
RF | 36.10 | 68.3 | 8 | 12.5 | 0.0423 | 145 | |
Taxi Trajectory | LTCOP-LSTM | 22.45 | 82.7 | 10 | 16.2 | 0.0188 | 105 |
ARIMA | 29.15 | 66.1 | 5 | 10.7 | 0.0506 | 158 | |
RF | 31.97 | 63.7 | 7 | 11.8 | 0.0395 | 140 | |
ULTRA | LTCOP-LSTM | 26.88 | 79.3 | 9 | 14.5 | 0.0214 | 115 |
ARIMA | 33.61 | 71.8 | 6 | 12.0 | 0.0467 | 152 | |
RF | 34.75 | 69.5 | 8 | 12.9 | 0.0402 | 142 |
