Data-driven Bayesian analysis of marine accidents in the English Channel

Abstract

The English Channel, renowned for its heavy traffic and challenging conditions, frequently witnesses marine accidents. To understand the causal mechanisms and identify risk factors, this study analyzed 833 accidents from 2017 to 2022, employing a Bayesian network predictive model. Validation using log loss, quadratic loss, and spherical gain metrics confirmed the model’s accuracy. The research elucidates the primary influences on accident occurrence: ship type, tonnage, accident location, and flag state. Ship-related factors surpass environmental ones in significance. Older ships face heightened risks of mechanical failures and collisions, especially in adverse weather with limited visibility. Yachts and fishing vessels are particularly susceptible to fire, explosion, collision, and grounding incidents. The study offers theoretical guidance for management departments to devise prevention strategies and enhance navigation safety in the English Channel.

Keywords:

• English Channel strait
• marine accidents
• data-driven Bayesian network
• predictive modeling
• causal analysis

Acknowledgments

The authors would like to thank the support from the Shanghai Philosophy and Social Science Planning Project (Grant No.2020BGL036). The authors also acknowledge the anonymous reviewers for their suggestions that improved the manuscript.

Authors’ contributions

Conceptualization, X. G. and Q. Y.; methodology, X. G. and Q. Y.; software, Y. W.; validation, Y. W.; formal analysis, Y. W., Q. Y., and X. G.; investigation, Y. W. and W. D.; resources, X. G.; data curation, Y. W.; writing—original draft preparation, Y. W. and W. D.; writing—review and editing, X. G. and Q. Y.; visualization, Y. W.; supervision, X. G.; project administration, X. G.; funding acquisition, X. G. All authors have read and agreed to the published version of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data presented in this study are available on request from the corresponding author.

Additional information

Funding

This research was funded by the Shanghai Office of Philosophy and Social Science, grant number 2020BGL036.