Road crashes have emerged as a major global health issue, particularly impacting vulnerable road users such as pedestrians, cyclists, and two-wheeler riders in developing nations. Existing traffic systems rely on classical learning models that are inefficient, less accurate, and limited to manual record-keeping without performing intelligent analysis. This paper presents a web-based real-time application integrated with data mining and unsupervised machine learning classification algorithms to analyze traffic crash data and predict the environmental, behavioral, and situational factors contributing to accidents. The proposed system automates pattern discovery and parameter tuning to discover hidden traffic associations, providing data-driven insights that assist traffic departments in implementing preventive road safety measures.
M et al. (Mon,) studied this question.