摘要
为厘清轨道交通运营风险中碎片化、口语化事件信息的演化逻辑,以轨道交通运营风险事件新闻中的部分内容作为数据源,抽取事件信息和事件关系,通过Neo4j构建轨道交通运营风险事件的事理图谱,进而利用Bert模型、层次聚类和莱文斯坦距离对风险事件进行泛化,并进一步梳理得出泛化事件的逻辑关系,由此得到轨道交通运营风险事件的演化路径和发展方向。研究结果表明:层次聚类方法中轮廓系数大于0.9,本文方法具有可行性。研究结果可为轨道交通运营风险防范及应急决策选择提供参考和支持。
In order to clarify the evolutionary logic of fragmented and colloquial event information in the operational risk of rail transit,taking part of the contents of news about the operation risk events in rail transit as the data source,the event information and event relationship were extracted,and an event causal graph of the operation risk events in rail transit was constructed by Neo4j.The Bert model,hierarchical clustering,and Levenstein distance were used to generalize the risk events.Furthermore,the logic relationship of generalized events was further clarified,and the evolution path and development direction of operational risk events in rail transit were obtained.The results show that the silhouette coefficient in the hierarchical clustering method is greater than 0.9,which demonstrates the feasibility of the proposed method.The research results can provide reference and support for risk prevention and emergency decision-making in rail transit operation.
作者
许慧
蔡林
XU Hui;CAI Lin(School of Economics and Management,Chongqing University of Posts and Telecommunications,Chongqing 400065,China;College of Science,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《中国安全生产科学技术》
CAS
CSCD
北大核心
2024年第4期12-19,共8页
Journal of Safety Science and Technology
基金
国家社会科学基金项目(22XGL013)。
关键词
轨道交通运营风险事件
事理图谱
Bert模型
层次聚类
事件泛化
operational risk events in rail transit
event causal graph
Bert model
hierarchical clustering
event generalization