面向港口突发事件的多层网络知识图谱构建  

Construction of Multi-layer Network Knowledge Graph for Port Emergency Events

作　　者：尹秀腾 李响[1] 刘国辉 芦鹏飞 张卫东 YIN Xiuteng;LI Xiang;LIU Guohui;LU Pengfei;ZHANG Weidong(Information Engineering University,Zhengzhou 450001,China;China Navigation Book Publishing House,Tianjin 300450,China;Unit 31016,Beijing 100081,China)

机构地区：[1]信息工程大学,河南郑州450001 [2]中国航海图书出版社,天津300450 [3]31016部队,北京100081

出　　处：《信息工程大学学报》2025年第2期189-195,共7页Journal of Information Engineering University

摘　　要：为克服传统研究中忽视相关要素层次性和关联性的问题,提出一种基于多层网络的港口突发事件知识图谱构建方法。该方法通过概念框架构建和实体要素映射,使用5W1H分析法和网络本体语言(OWL)构建事件本体和港口本体,并基于一种多层网络模型对实体间关系进行建模。以天津港爆炸事故为例,应用协议和资源描述框架查询语言(SPARQL)进行查询实验分析。结果表明,该方法能够实现高效的基于关系的数据查询,减少多跳查询中的循环结果问题,提升查询结果的准确性,有助于深入理解港口突发事件对港口多维度要素的跨领域影响。To address the limitations of hierarchical structure oversight and interconnected element neglect in conventional research,a multi-layer network-based methodology for constructing port emergency knowledge graphs is proposed.Event and port ontologies are systematically constructed through conceptual framework development and entity element mapping,with the 5W1H analytical method and web ontology language(OWL)being applied.Based on the ontologies,inter-entity relationships are modeled using a kind of multi-layer network architecture.The Tianjin port explosion incident is examined as a case study,with the SPARQL protocol and RDF query language(SPARQL)being implemented for experimental query analysis.It is demonstrated that this approach enables efficient relationship-driven data retrieval while mitigating circular result generation in multi-hop queries,and achieve enhanced query accuracy.The method will help to gain to a deeper understanding of cross-domain impacts of port emergencies on multidimensional port elements.

关 键 词：港口突发事件 多层网络 知识图谱 5W1H分析法 通用事件模型 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程] U658[自动化与计算机技术—控制科学与工程]