Spatiotemporal evolution of global maritime accidents: Integrating hot spot detection and severity modeling for system safety
提出一个结合时空热点检测和逻辑回归的框架，分析全球海事事故的时空演化趋势，识别事故热点水域，并探究事故类型、船舶状况等因素与严重性的关系，为航运安全管理提供指导。
As an important pillar of international trade, the shipping industry has become increasingly important in the global economy. However, the frequent occurrence of maritime transportation accidents has posed a threat to the human life safety and marine environment as well. In this study, we propose a novel integrated framework that combines spatiotemporal hotspot detection and severity-oriented risk modeling that combines spatial density analysis and emerging spatio-temporal hot spot analysis to investigate the evolutionary trend of global maritime accidents from both temporal and spatial dimensions, identifying accident hot spot waters, and analyzing the relationship between influencing factors (e.g., type of accident, type of vessel, and condition of ships) and the severity of accidents by using logistic regression models. The results indicate that the spatial distribution of maritime accidents has apparent hot spot agglomeration characteristics of dynamic evolutionary trends. The accident hot spot areas show significant changes in different time periods, which are mainly concentrated in the shipping-intensive areas. Similar trends are also seen in other shipping hub regions such as northwestern Europe, eastern North America and northwestern Africa. The study provides an important theoretical basis and practical guidance for the development of shipping safety management and accident prevention measures, which can help reduce the occurrence of maritime accidents and their severity.