Content analysis of marine incident reports: The causal factors and associative relationship

Marine incident reports available from official marine bodies contain a wealth of information about the type, causes of incident, and details on the sequence of events that lead to the incident. Knowledge discovery and retrieval of useful information through content analysis and text mining from such reports will help us to understand the primary/secondary causes for the incidents and establish the associative relationship between the causal factors and the type of incidents. In this work, we used the marine incident reports obtained from the `National Transportation Safety Board, US'. These reports were classified as per the incident type and the primary/secondary causes of the incidents were identified. In addition, heat map(s) to establish the relationship between the primary/secondary causes with respect to the incident type were generated. Causal factors and associative relationships that were not apparent in earlier studies were also identified using the proposed methodology. Additional information on the incidents was also collected to list down the lessons learnt from the historical events so as to formulate the recommendations for mariners to aid their future navigation.