Bayesian network modeling of Port State Control inspection findings and ship accident involvement
作者：
Highlights：
• Bayesian network models of Port State Control findings and accidents are proposed.
• NPC algorithm outperforms score-based greedy search in learning the model.
• A hidden variable representing the complete ship system and its safety is introduced.
• Models with a binary system safety variable is preferred over the other alternatives.
• Structural conditions is the most informative deficiency type for accident probability.
摘要
•Bayesian network models of Port State Control findings and accidents are proposed.•NPC algorithm outperforms score-based greedy search in learning the model.•A hidden variable representing the complete ship system and its safety is introduced.•Models with a binary system safety variable is preferred over the other alternatives.•Structural conditions is the most informative deficiency type for accident probability.
论文关键词：Maritime traffic accidents,Bayesian networks,Port State Control,Hidden variables,Maritime traffic safety
论文评审过程：Available online 2 September 2013.
论文官网地址：https://doi.org/10.1016/j.eswa.2013.08.060
