Unraveling the patterns of critical contributory factors and flag-state affiliations linked with maritime incident outcomes

Abstract

Maritime incidents are complex in nature and result from various factors. Although researchers have explored the factors influencing maritime safety, it is unclear how the critical contributory factor chains including flag-state associations, influence maritime incident outcomes, namely deaths, injury, and missing person cases. This paper applies a robust non-parametric Association Rules Mining (ARM) algorithm to an extensive marine incident database to identify critical patterns of key contributory factors and flag-state associations leading to different maritime incident casualty outcomes. The study dataset comprised 13,758 maritime incidents recorded between 2002 and 2015 in US territorial waters. This data was sourced from the Marine Information for Safety and Law Enforcement (MISLE) database system and served as the basis for analysis. A preliminary investigation revealed that low-weight passenger ships were most likely involved in maritime incidents. Further analysis established that, whereas Bahamas-registered vessels were highly likely to be involved in incidents associated with deaths, US-registered vessels were primarily related to injury-related incidents. Singaporean ships were often involved in missing-person incidents. Additionally, the results highlighted significant associations between ocean cruise vessels and existing medical conditions with deaths. Besides, passengers, towing service vessels, and crewmembers were identified to be highly associated with injury. Higher-order associations depicted that tank ships, fishing vessels, ship masters and crew members are likely to be involved in maritime incidents where missing persons are involved. Policies directed at eliminating/controlling the contributory factors identified to lead to a particular outcome to help reduce the occurrence or impacts of maritime incidents have been suggested.