Papers › Pattern-Matching Dynamic Memory Network for Dual-Mode Traffic Prediction
Pattern-Matching Dynamic Memory Network for Dual-Mode Traffic Prediction
Wenchao Weng, Mei Wu, Hanyu Jiang, Wanzeng Kong, Xiangjie Kong, Feng Xia
In recent years, deep learning has increasingly gained attention in the field of traffic prediction. Existing traffic prediction models often rely on GCNs or attention mechanisms with O(N^2) complexity to dynamically extract traffic node features, which lack efficiency and are not lightweight. Additionally, these models typically only utilize historical data for prediction, without considering the impact of the target information on the prediction. To address these issues, we propose a Pattern-Matching Dynamic Memory Network (PM-DMNet). PM-DMNet employs a novel dynamic memory network to capture traffic pattern features with only O(N) complexity, significantly reducing computational overhead while achieving excellent performance. The PM-DMNet also introduces two prediction methods: Recursive Multi-step Prediction (RMP) and Parallel Multi-step Prediction (PMP), which leverage the time features of the prediction targets to assist in the forecasting process. Furthermore, a transfer attention mechanism is integrated into PMP, transforming historical data features to better align with the predicted target states, thereby capturing trend changes more accurately and reducing errors. Extensive experiments demonstrate the superiority of the proposed model over existing benchmarks. The source codes are available at: https://github.com/wengwenchao123/PM-DMNet.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Traffic Prediction | PeMS07 | PM-DMnet(R) | MAE@1h | 19.18 | #3 of 17 | Archive leaderboard | report |
| Traffic Prediction | PeMS07 | PM-DMNet(P) | MAE@1h | 19.35 | #5 of 17 | Archive leaderboard | report |
| Traffic Prediction | PeMS08 | PM-DMNet(R) | MAE@1h | 13.40 | #3 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMS08 | PM-DMNet(P) | MAE@1h | 13.55 | #7 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD4 | PM-DMNet(P) | 12 steps MAE | 18.34 | #6 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD4 | PM-DMNet(P) | 12 steps MAPE | 12.05 | #6 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD4 | PM-DMNet(P) | 12 steps RMSE | 30.36 | #6 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD4 | PM-DMNet(R) | 12 steps MAE | 18.37 | #7 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD4 | PM-DMNet(R) | 12 steps MAPE | 12.01 | #7 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD4 | PM-DMNet(R) | 12 steps RMSE | 30.68 | #7 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7 | PM-DMNet(R) | 12 steps MAE | 19.18 | #3 of 8 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7 | PM-DMNet(R) | 12 steps MAPE | 7.95 | #3 of 8 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7 | PM-DMNet(R) | 12 steps RMSE | 33.15 | #3 of 8 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7 | PM-DMNet(P) | 12 steps MAE | 19.35 | #4 of 8 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7 | PM-DMNet(P) | 12 steps MAPE | 8.05 | #4 of 8 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7 | PM-DMNet(P) | 12 steps RMSE | 33.33 | #4 of 8 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(L) | PM-DMNet(R) | 12 steps MAE | 2.79 | #3 of 6 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(L) | PM-DMNet(R) | 12 steps MAPE | 6.99 | #3 of 6 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(L) | PM-DMNet(R) | 12 steps RMSE | 5.81 | #3 of 6 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(L) | PM-DMNet(P) | 12 steps MAE | 2.81 | #4 of 6 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(L) | PM-DMNet(P) | 12 steps MAPE | 7.13 | #4 of 6 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(L) | PM-DMNet(P) | 12 steps RMSE | 5.79 | #4 of 6 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(M) | PM-DMNet(R) | 12 steps MAE | 2.60 | #3 of 7 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(M) | PM-DMNet(R) | 12 steps MAPE | 6.57 | #3 of 7 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(M) | PM-DMNet(R) | 12 steps RMSE | 5.36 | #3 of 7 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(M) | PM-DMNet(P) | 12 steps MAE | 2.61 | #4 of 7 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(M) | PM-DMNet(P) | 12 steps MAPE | 6.55 | #4 of 7 | Archive leaderboard | report |
| Traffic Prediction | PeMSD7(M) | PM-DMNet(P) | 12 steps RMSE | 5.33 | #4 of 7 | Archive leaderboard | report |
| Traffic Prediction | PeMSD8 | PM-DMNet(R) | 12 steps MAE | 13.40 | #3 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD8 | PM-DMNet(R) | 12 steps MAPE | 8.87 | #3 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD8 | PM-DMNet(R) | 12 steps RMSE | 23.22 | #3 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD8 | PM-DMNet(P) | 12 steps MAE | 13.55 | #6 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD8 | PM-DMNet(P) | 12 steps MAPE | 9.04 | #6 of 13 | Archive leaderboard | report |
| Traffic Prediction | PeMSD8 | PM-DMNet(P) | 12 steps RMSE | 23.35 | #6 of 13 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections