同济大学学报（自然科学版）2025，Vol.53Issue(4)：611-618,8.DOI:10.11908/j.issn.0253-374x.23307
交通事故致因知识图谱构建及风险因素挖掘
Traffic Accident Causation Knowledge Graph Construction and Risk Factor Mining
摘要
Abstract
In this paper,we use data from traffic accident investigation reports to construct a traffic accident causation knowledge graph and analyze risk factors.Firstly,we construct the recognition model of named entities of traffic accident causation applicable to low data volume based on the fine-tuned UIE pre-training model for the generation of the entity set.Secondly,through the structured processing and ontology construction,the graph database Neo4j is used to store the traffic accident causation knowledge graph for visualization.Thirdly,based on the expert experience and pre-trained language text classification model,the traffic accident causation entities are standardized.Finally,a risk factor analysis method based on the traffic accident causation graph is constructed to mine triggering characteristics and contributions of each factor by analyzing the type distribution and degree distribution of standardized entities,and to perform the association rule mining.The results of these methods and analyses provide an in-depth understanding and exploration of historical accident risk factors.关键词
交通运输/知识图谱/致因分析/数据挖掘/命名实体识别Key words
transportation/knowledge graph/causal analysis/data mining/named entity recognition分类
交通工程引用本文复制引用
王占中,张书源,杨萌,兰若冰,吴智豪..交通事故致因知识图谱构建及风险因素挖掘[J].同济大学学报（自然科学版）,2025,53(4):611-618,8.基金项目
吉林省自然科学基金面上项目(20230101112JC) （20230101112JC）
