Table 1 Summarization of existing methods.
Reference | Work | Methodology | Advantages | Drawbacks |
|---|---|---|---|---|
Traffic prediction model | LT-GCN | • Minimize training time | • Did not focus on the false positive rate | |
Interpretable traffic prediction | Traffexplainer | • Improved prediction performance • minimal root mean square error | • Compromised training time | |
Vehicle flow and vehicle speed predictions | Physics-aware recurrent neural network | • Precise prediction | • The time factor was not analyzed | |
Traffic prediction | Knowledge-sequence-to-sequence (K-Seq2Seq) | • Traffic prediction with minimal error | • Lack optimal traffic management | |
Traffic prediction | Deep learning | • High prediction accuracy | • High root mean square error | |
Accurate network management | Explainable deep learning | • Efficient resource allocation | • High training time | |
Forecasting traffic prediction | Long-short-combination (LSC) | • Improved forecasting accuracy | • Increased root mean square error | |
Traffic prediction in the urban road network | Multi-graph Learning | • Enhanced prediction accuracy | • High prediction time | |
Smart city traffic prediction | Hybrid gated recurrent with LSTM | • Improved accuracy | • Low precision and training time | |
Traffic congestion handling | Convolutional neural and LSTM networks | • Minimized traffic congestion time | • Compromised accuracy rate |
