"Uncovering Influential Mechanisms of Traffic Accident Severity through Explainable Machine Learning"

"This study utilizes the UK STATS19 road accident dataset, which provides detailed records of personal injury crashes from 2005 to 2014. The dataset integrates multi-source information, including traffic conditions, roadway geometry, environmental factors, vehicle characteristics, and casualty attributes. With its large scale, structured format, and encoded variables, it offers a robust foundation for analyzing accident occurrence and severity patterns. The data enable the exploration of spatiotemporal distributions, key risk factors, and complex interaction effects influencing crash outcomes. Its richness and high dimensionality make it particularly suitable for machine learning-based prediction, interpretable modeling, and data-driven traffic safety analysis, supporting more effective decision-making in road safety management and infrastructure planning."