Towards Large-Scale Graph-Based Traffic Forecasting: A Data-Driven Network Partitioning Approach
- Publisher:
- Institute of Electrical and Electronics Engineers
- Publication Type:
- Journal Article
- Citation:
- IEEE Internet of Things Journal, 2022, 10, (5), pp. 4506-4519
- Issue Date:
- 2022-01-01
Closed Access
| Filename | Description | Size | |||
|---|---|---|---|---|---|
| Toward_Large-Scale_Graph-Based_Traffic_Forecasting_A_Data-Driven_Network_Partitioning_Approach.pdf | Published version | 11.12 MB |
Copyright Clearance Process
- Recently Added
- In Progress
- Closed Access
This item is closed access and not available.
Network partitioning is recognized as an effective auxiliary approach for solving transportation tasks on large-scale traffic networks in a domain-decomposition manner. Most of the existing related partitioning algorithms are explicitly designed to traffic management problems and merely focus on the implied topology of the networks. In this paper, towards the practical problems that happened to traffic forecasting tasks, we propose a network-partitioning-based domain-decomposition framework to improve GCN-based predictors’ performance on large-scale transportation networks. Particularly, we devise a data-driven network-partitioning approach, namely, Speed-Matching-Partitioning, which employs not only the topological features but also the traffic speed observations of traffic networks for partitioning. Additionally, we propose a data-parallel training strategy that feeds partitioned sub-networks into independent predictors for parallel training. The proposed approach is tested by comprehensive case studies on three real-world datasets to evaluate its effectiveness. The results indicate that the proposed approach can help improve GCN-based predictors’ accuracy and training efficiency on both small and relatively large traffic datasets. Furthermore, we investigate the model sensitivity to the selection of graph representations and framework parameters, and the learning efficiency of the data-parallel training strategy.
Please use this identifier to cite or link to this item: