A spatial graph learning framework for multi‐scale road safety management based on road‐curve features and open‐source data
Bibliographic record
Abstract
Horizontal and vertical curves significantly affect crash risk due to their impact on driver behavior, vehicle dynamics, and sight distance. However, their combined effects and spatial interactions remain underexplored in large-scale safety assessments. To address limitations in high-resolution geometric data and insufficient spatial modeling, this study proposes a geometry-oriented crash risk assessment framework based on graph neural networks. Leveraging open-source geospatial data, this study extracts fine-grained curve features and constructs a GraphSAGE model to capture spatial dependencies among road segments. A dual-graph architecture is developed to jointly encode both segment-level and network-level information. In large-scale empirical evaluations, the proposed model exhibits excellent predictive performance (F1 > 0.985) and strong spatial correlation with historical crash distributions (r > 0.7). The model effectively identifies high-risk segments characterized by poor geometric continuity or abrupt structural transitions, providing decision support for alignment optimization. The model effectively identifies high-risk segments characterized by poor geometric continuity or abrupt structural transitions, thereby supporting informed decisions for alignment improvements. This research enhances the understanding of the geometry–safety relationship and offers a scalable, open-source tool to support local and regional traffic safety interventions.
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 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".