摘要
为准确分析出影响海上船舶安全的隐患类别,提高船舶、船员、设备的管理制度,提出一种基于双向编码器(BERT)的海上船舶安全隐患分类。采用Text Rank算法对隐患文本进行关键词提取,将提取的关键词输入到BERT预训练模型,提升模型在分类任务上的性能。试验结果表明,训练前采用Text Rank算法进行关键词提取,准确率、召回率和F1值都有所提升,验证了模型的有效性,达到了快速准确对海上船舶安全隐患分类的目的。
In order to accurately analyze the types of hidden dangers affecting the safety of marine vessel and improve the management system of ships,crew and equipment,a classification of hidden dangers of ships at sea based on bidirectional encoder representation from transformers(BERT)is proposed.TextRank algorithm was used to extract keywords from hidden text and input the extracted keywords into BERT pre-training model to improve the performance of the model on classification tasks.The experimental results show that the accuracy,recall rate and F1 value of keywords extracted by TextRank algorithm before training have been improved,which verifies the effectiveness of the model and achieves the purpose of fast and accurate classification of maritime vessel safety hazards.
作者
靳嵩
朱艳
吴可嘉
孟祥松
赵乾菊
王颖
JIN Song;ZHU Yan;WU KeJia;MENG Xiangsong;ZHAO Qianju;WANG Ying(External Cooperation Program Department of Petro China Dagang Oilfield Company,Tianjin 300280,China;China Petroleum Pipeline Engineering Corporation Tianjin Branch,Tianjin 300450,China)
出处
《船舶工程》
CSCD
北大核心
2023年第S01期381-384,共4页
Ship Engineering
