Answers are grounded in this paper and cite their sources.
Analysis identifies accident-prone areas by integrating spatial and data mining methods, suggesting improvements for traffic safety.
This study presents a holistic approach to traffic accident analysis by integrating hierarchical clustering, variogram modelling and association rule mining. The analysis identified four critical accident-prone zones: dense residential areas, roads near city centres, multi-curved roads, and dispersed residential and agricultural areas. The Gaussian variogram model revealed significant spatial dependencies, indicating that accidents are concentrated in specific hotspots rather than evenly distributed. Association rule mining revealed key factors contributing to accidents, including dry road surfaces, fair weather conditions and the absence of public transportation vehicles. Additionally, road geometry, particularly overlapping horizontal and vertical curves, significantly contributed to accident frequency in multi-curved regions. The study’s findings align with existing literature, offering deeper insights through a unified framework. Recommendations include the implementation of advanced traffic monitoring systems, improvements in road infrastructure and targeted driver education, contributing to more effective traffic safety strategies and accident prevention.
No takes yet. Share an insight, caveat, or question.
Cansiz et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: