Identification of Urban Traffic Accident Hotspots Using Getis-Ord Gi and DBSCAN Algorithmsimes

This study aims to identify urban traffic accident hotspots by integrating spatial statistics and clustering techniques. Initially, the Getis-Ord Gi* method is employed to preliminarily identify areas with high accident frequencies. Subsequently, the DBSCAN algorithm is applied for further analysis of these hotspot areas. To optimize the two key parameters of DBSCAN (eps and min_samples), this study utilizes GridSearchCV for parameter search and adopts the adjusted silhouette score as the evaluation criterion. The results indicate that the proposed method effectively overcomes the limitations of traditional approaches, enhancing the accuracy of traffic accident hotspot identification.