Tag
The paper proposes a retrieval-augmented generation framework using large language models to recommend safety countermeasures from crash narratives, achieving high precision and recall in evaluation on real-world data.
This paper presents a schema-grounded natural language interface for transportation safety analysis that uses a large language model to interpret user queries while preserving deterministic execution against an authoritative database. The framework is evaluated on a Massachusetts transportation safety database, successfully executing all queries and correcting errors in 29% of cases, demonstrating a practical approach to broadening access to safety data.
This study presents a hybrid predictive framework using CatBoost and SHAP to identify risk factors in tree-involved traffic crashes, highlighting restraint non-use as the most critical predictor of severe injury.