How Do Navy Escorts Influence Piracy Risk in East Africa? A Bayesian Network Approach.
Risk analysis(2024)
Abstract
Navy escorts are considered crucial in countering illegal piracy attacks. In this paper, a novel approach is developed to investigate the effect of navy escorts on piracy incidents by models based on two enhanced Tree-Augmented Naïve (TAN) Bayesian networks. This approach offers a systematic investigation into the various factors that influence pirate activities, and helps to identify changes in piracy attack behaviors when confronted by navy escorts and assess the effectiveness of anti-piracy measures. An empirical study is conducted utilizing a unique data set compiled from multiple sources from 2000 to 2019. The empirical evidence shows that there was a gradual reduction in the incidence of piracy attacks in East Africa following the implementation of navy escorts in 2009, but with a surge in 2010 and 2011. The data set is, thus, divided into two time periods at the point of 2009 to facilitate a robust and comprehensive analysis, resulting in the development of two TAN models. Meanwhile, the geographical distribution of pirate attacks has shifted from international waters to port areas and territorial waters. We argue that the surge and geographical shift could be attributed to the calculating behavior of pirates when they encounter external pressures. Finally, a Shapely approach is introduced to evaluate the potential effectiveness of the implemented risk management strategies from a Game Theory perspective. This study offers new insights into the promotion of navy escorts and contributes to the development of a framework for assessing piracy risks in uncertain and dynamic anti-piracy environments.
MoreTranslated text
Key words
Bayesian network,maritime piracy,maritime safety,navy escorts,risk assessment
AI Read Science
Must-Reading Tree
Example
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined