Graph partitioning for accuracy and scalability in training GNN-based traffic prediction models for Intelligent Transportation Systems
Bibliographic record
Abstract
As urban areas grow and transportation networks become increasingly complex, accurate large-scale traffic forecasting is essential for Intelligent Transportation Systems. Graph Neural Networks-based traffic prediction models, such as the Diffusion Convolutional Recurrent Neural Networks (DCRNN) model, treat the traffic flow as a diffusion process on the graph, effectively capturing the propagation of the traffic conditions through the network. Unlike linear statistical models, GNNs inherently capture non-Euclidean spatial dependencies via graph structures. However, training GNNs on large-scale graphs is resource-intensive and time-consuming. This paper aims to study the impact of graph partitioning tools on accuracy and scalability in training GNN-based traffic prediction models. This paper integrates the Adaptive Graph Convolutional Recurrent Network (AGCRN) with a domain-specific graph partitioning tool for transportation networks—the Buffoon-optimized Karlsruhe High-Quality Partitioning (KaHIP). The partitions are trained on an NVIDIA A100 Tensor Core GPU. Tests on the California highway network dataset show that the Buffoon-optimized KaHIP framework produces higher-quality partitions, resulting in higher forecasting accuracy. We also show that the AGCRN-KaHIP integration outperforms DCRNN-METIS and baseline statistical methods. The integration achieves lower prediction errors and more stable forecasts.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".