Abstract—This study proposes a context-aware approach for risk ranking of railway level crossings using machine learning and national transport data. Railway level crossings are critical safety points within multimodal transportation systems, where interactions between rail operations and road traffic generate heterogeneous risk. Conventional risk assessment methods often rely on fixed-weight or expert-driven scoring schemes, which may not adequately capture context-dependent variations across diverse infrastructural, road, and spatial environments. To address these limitations, the proposed approach integrates national Geographic Information System (GIS)-based railway crossing data with machine learning and weakly supervised learning techniques. Contextual information is organized into three dimensions: infrastructural attributes, road characteristics, and spatial context. In the absence of explicit accident labels, pseudo-labels derived from domain knowledge are used to provide supervisory signals for model training. Multiple machine learning models, including logistic regression, random forest, and gradient boosting, are evaluated using ranking-oriented performance metrics to emphasize relative risk discrimination rather than binary accident prediction. The results indicate that the proposed approach generates continuous and discriminative risk scores, revealing substantial heterogeneity in relative safety risk across the railway network. A relatively small subset of crossings accounts for a disproportionate share of elevated risk, supporting targeted prioritization. Overall, the findings suggest that context-aware, data-driven risk ranking provides a scalable and interpretable alternative to traditional fixed-weight assessment methods, contributing to improved analytical understanding of safety risk patterns in railway level crossing networks.
 
Keywords—railway level crossings, context-aware risk scoring, transportation safety, machine learning, weakly supervised learning, risk ranking, Geographic Information System (GIS)-based transport data, safety prioritization
 
Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).