海事事故严重性影响因素及影响程度识别

来源：交通运输系统工程与信息杂志2019年第1期北京时间：

作者：吴琴，施欣，陶学宗

单位：上海海事大学 交通运输学院，上海 201306

摘要：Based on the investigation reports of maritime traffic accident, which were published by three maritime safety administrations, Shanghai, Zhejiang and Jiangsu, two kinds of ordered probability model (including the ordered Logit and the ordered Probit model) are established to identify the influencing factors and their impact levels on the severity of maritime traffic accidents (SMTA). The results show that the ordered probabilistic models are suitable for the study, but the ordered Logit model outperforms the ordered Probit model. In particular, the factors of season, time interval, weather, navigable waterways, vessel type, length and ownership have positive influences on larger and above accidents, while the influence of visibility is negative. Among them, the impact of the vessel length above 200 meters is the largest (marginal effect: 0.378 3), which followed by summer season (0.282 2). Meanwhile, the impact of poor visibility is smallest (- 0.108 0), which followed by individual vessels (0.109 5). Finally, the results can provide a decision basis for maritime safety administrations to carry out safety supervision work.

关键词：waterwaytransportation,impactlevel,orderedprobabilitymodel,severityofmaritimetrafficaccidents,marginaleffectanalysis

基金资助：国家自然科学基金/National Natural Science Foundation of China(71573172).