Construction and Application of Marine Accidents Knowledge Graph

Abstract:
Marine accident reports contain a wealth of knowledge that has not been fully explored and utilized in existing studies. To extract the implied important information from a large number of marine accidents and effectively manage the marine accidents knowledge, this paper takes the local marine accident reports issued by China Maritime Safety Administration (MSA) as the data source, and proposes to construct the marine accidents knowledge graph through the combination of top-down and bottom-up methods. Firstly, the model layer of the knowledge graph is defined. Then the related entities are extracted by using the named entity identification technology based on BERT-BiLSTM-CRF and stored in the Neo4j graph database. Then, the knowledge graph of marine accidents is preliminarily constructed based on experiments to verify the effectiveness of the above methods. On this basis, the application of the marine accident knowledge graph is analyzed.