Maritime security threats: Classifying and associating patterns in piracy and armed robbery incidents

Abstract

Piracy and armed robbery (P&AR) incidents are one of the most significant security problems for the maritime industry. These incidents, which negatively affect maritime activities and cause disruptions in the global supply chain, continue in many different parts of the world. This study focuses on the Gulf of Guinea, the region with the highest occurrence of P&AR today. The study begins with data pre-processing, applied to a total of 1076 P&AR reports from the Gulf of Guinea, sourced from the Global Integrated Shipping Information System (GISIS) database. Various machine learning (ML) algorithms are then utilized to determine the best-performing model for imputing missing data. Next, the Chi-square Automatic Interaction Detection (CHAID) algorithm is employed to classify P&AR incidents. Finally, the Apriori algorithm, a method from Association Rule Mining (ARM), is used to uncover hidden relationships and associations within the dataset. Additionally, the findings are visualised to make interpreting the results more accessible and transparent. The analysis results reveal that weapons, coastal authority, and ship size have a significant impact on the occurrence of attacks. Robbery attacks typically target storerooms using knives during night port activities. In contrast, kidnapping incidents involve armed attackers directly targeting the accommodation areas of ships with low tonnage and speed. In hijacking incidents, large groups of attackers operate in international waters, primarily targeting tanker ships aged over 12 years with low freeboard. In conclusion, the findings of this study aim to assist authorities and ship operating in the region in implementing necessary precautions.