基于事理图谱相似度的高速公路交通事故持续时间预测

doi:10.16265/j.cnki.issn1003-3033.2026.07.1372

摘要/Abstract

摘要：

为缓解高速公路交通事故造成的拥堵问题、提升道路通行效率,提出基于事理图谱相似度(S-EEG)的高速公路事故持续时间预测模型。首先,构建涵盖属性、实体、事件和关系的高速公路交通事故本体模型,运用事理图谱技术,融合结构化数据与文本数据的事理逻辑知识,揭示事故演化规律与应急处置特征;然后,构建基于最大公共子图(MCS)的图结构相似性评价模型,给出超参数K值和统计推断的寻优策略;最后,经知识抽取与事件泛化构建抽象事理图谱,以K个高相似度事故案例持续时间的统计推断实现预测,并利用陕西省高速公路交通事故记录进行实证分析,构建包含242 150个节点和251 596条有向边的高速公路交通事故事理图谱。结果表明:预测模型的平均绝对百分比误差(MAPE)为42.33%,较基准模型降低56.68%;在交通事故持续时间的分类任务中,性能提升约2.31%。

关键词: 事理图谱相似度(S-EEG), 高速公路, 交通事故, 持续时间预测, 本体模型, 最大公共子图(MCS)

Abstract:

To mitigate traffic congestion resulting from expressway accidents and enhance overall operational efficiency, this paper proposes a S-EEG method for predicting expressway traffic accident duration based on the similarity of event evolution graphs. Initially, an ontology model of expressway traffic accidents was constructed using attributes, entities, events, and relationships. The event evolution graph technology was employed to integrate logical knowledge from structured data and textual data, revealing the evolution patterns of traffic accidents and characteristics of emergency response. Subsequently, a graph structure similarity evaluation model based on MCS was proposed, along with an optimization method for determining the hyperparameter K and statistical parameter. Finally, an abstract event evolution graph for expressway traffic accidents was built through knowledge extraction and event generalization, and the prediction of traffic accident duration is achieved by statistically inferring the duration of K highly similar cases. Furthermore, an empirical analysis was carried out using expressway traffic accident records from Shaanxi Province, and a corresponding event evolutionary graph for expressway traffic accidents was built, comprising 242,150 nodes and 251,596 directed edges. The results indicate that the proposed model achieves a mean absolute percentage error (MAPE) of 42.33%, reducing the prediction error by 56.68% compared to the baseline model. Additionally, it improves performance by approximately 2.31% in the classification task of accident duration.

中图分类号:

X951交通运输安全

陈娇娜, 张瑾, 毛依文, 靳引利. 基于事理图谱相似度的高速公路交通事故持续时间预测[J]. 中国安全科学学报, 2026, 36(7): 77-85.

Chen Jiaona, Zhang Jin, Mao Yiwen, Jin Yinli. Research on duration prediction of traffic accident on expressways based on similarity of event evolution graphs[J]. China Safety Science Journal, 2026, 36(7): 77-85.

图/表 14

图1

图2

表1

图3

图4

图5

图6

图7

图8

表2

表3

表4

图9

图10

参考文献 32

[1]	世界观\|总里程突破19万公里, 中国编织全球最大高速公路网[EB/OL]. [2025-07-21]. https://news.youth.cn/jsxw/202507/t20250730_16147585.htm.
[2]	Li Ruimin, Pereira F C, Ben-akiva M E. Overview of traffic incident duration analysis and prediction[J]. European Transport Research Review, 2018, 10 (2): 1-13. doi: 10.1007/s12544-017-0273-5
[3]	丛浩哲, 方守恩, 王俊骅. 交通事件持续时间影响因素分析及其回归模型[J]. 交通信息与安全, 2010, 28 (3): 80-83.
	Cong Haozhe, Fang Shouen, Wang Junhua. Factors analysis of freeway incident duration and regression modeling[J]. Journal of Transport Information and Safety, 2010, 28(3): 80-83.
[4]	Wang Shi, Li Ruimin, Guo Min. Application of nonparametric regression in predicting traffic incident duration[J]. Transport, 2015, 33 (1): 22-31. doi: 10.3846/16484142.2015.1004104
[5]	Yu Bin, Wang Yitian, Yao Jinbao, et al. A comparison of the performance of ANN and SVM for the prediction of traffic accident duration[J]. Neural Network World, 2016, 26 (3): 271-287. doi: 10.14311/NNW.2016.26.015
[6]	Zhu Weiwei, Wu Jinglin, Fu Ting, et al. Dynamic prediction of traffic incident duration on urban expressways: a deep learning approach based on LSTM and MLP[J]. Journal of Intelligent and Connected Vehicles, 2021, 4 (2): 80-91. doi: 10.1108/JICV-03-2021-0004
[7]	Li Ruimin. Traffic incident duration analysis and prediction models based on the survival analysis approach[J]. IET Intelligent Transport Systems, 2015, 9 (4): 351-358. doi: 10.1049/itr2.v9.4
[8]	马雪婧, 邵春福, 钱剑培, 等. 交通事件持续时间预测的贝叶斯网络模型[J]. 交通信息与安全, 2015, 33 (6): 65-71.
	Ma Xuejing, Shao Chunfu, Qian Jianpei, et al. Bayesian network model for the prediction of traffic incident duration[J]. Journal of Transport Information and Safety, 2015, 33(6): 65-71.
[9]	Linchao Li, Xi Sheng, Bowen Du, et al. A deep fusion model based on restricted Boltzmann machines for traffic accident duration prediction[J]. Engineering Applications of Artificial Intelligence, 2020, 93: DOI:10.1016/j.engappai.2020.103686.
[10]	Ghosh B, Dauwels J. Comparison of different Bayesian methods for estimating error bars with incident duration prediction[J]. Journal of Intelligent Transportation Systems, 2022, 26(4):420-431. doi: 10.1080/15472450.2021.1894936
[11]	何珂, 杨顺新, 郜勇刚. 基于PCA-RF组合模型的隧道交通事故持续时间预测[J]. 交通信息与安全, 2019, 37(5): 26-32.
	He Ke, Yang Shunxin, Gao Yonggang. Prediction of traffic incident duration in tunnels based on a PCA-RF combined model[J]. Journal of Transport Information and Safety, 2019, 37(5): 26-32.
[12]	Shang Qiang, Tan Derong, Gao Song, et al. A hybrid method for traffic incident duration prediction using BOA-optimized random forest combined with neighborhood components analysis[J]. Journal of Advanced Transportation, 2019: DOI: 10.1155/2019/4202735.
[13]	Saracoglu A, Ozen H. Estimation of traffic incident duration: a comparative study of decision tree models[J]. Arabian Journal for Science and Engineering, 2020, 45 (10): 8 099-8 110.
[14]	Li Kuang, Han Yan, Zhu Yujia, et al. Predicting duration of traffic accidents based on cost-sensitive Bayesian network and weighted K-nearest neighbor[J]. Journal of Intelligent Transportation Systems, 2019, 23 (2): 161-174. doi: 10.1080/15472450.2018.1536978
[15]	Tang Jinjun, Zheng Lanlan, Han Chunyang, et al. Traffic incident clearance time prediction and influencing factor analysis using extreme gradient boosting model[J]. Journal of Advanced Transportation, 2020, 2020(1): DOI: 10.1155/2020/6401082.
[16]	吕路, 李杰, 郭忠印, 等. 高速公路交通事故持续时间计算方法研究[J]. 公路交通科技, 2022, 39 (12): 155-162. doi: 10.3969/j.issn.1002-0268.2022.12.019
	Lyu Lu, LI Jie, Guo Zhongyin, et al. Study on calculation method of expressway accident duration[J]. Journal of Highway and Transportation Research and Development, 2022, 39(12): 155-162.
[17]	Zhao Yuexu, Deng Wei. Prediction in traffic accident duration based on heterogeneous ensemble learning[J]. Applied Artificial Intelligence, 2022, 36(1): DOI:10.1080/08839514.2021.2018643.
[18]	Hamad K, Ruzouq A R, Zeiada W, et al. Predicting incident duration using random forests[J]. Transportmetrica A: Transport Science, 2020, 16 (3): 1269-1293. doi: 10.1080/23249935.2020.1733132
[19]	Grigorev A, Mihaita A S, Lee S, et al. Incident duration prediction using a bi-level machine learning framework with outlier removal and intra-extra joint optimization[J]. Trasportation Research Part C-emerging Technologies, 2022,141:DOI:10.1016/j.trc.2022.103721.
[20]	Zhao Hui, Gunardi W, Liu Yang, et al. Prediction of traffic incident duration using clustering-based ensemble learning method[J]. Journal of Transportation Engineering, Part A: Systems, 2022, 148(7):DOI:10.1061/JTEPBS.0000688.
[21]	贾兴利, 李双庆, 杨宏志, 等. 基于ATT-LSTM模型的高速公路交通事件持续时长预测[J]. 交通信息与安全, 2022, 40 (5): 61-69.
	Jia Xingli, Li Shuangqing, Yang Hongzhi, et al. Prediction of the duration of freeway traffic incidents based on an ATT-LSTM model[J]. 交通信息与安全, 2022, 40(5): 61-69.
[22]	何庆龄, 刘静, 李珊, 等. 基于SO-BiLSTM的高速公路交通事故持续时间预测[J]. 重庆交通大学学报(自然科学版), 2024, 43 (10): 97-105.
	He Qingling, Liu Jing, Li Shan, et al. Highway traffic accident duration prediction based on SO-BiLSTM[J]. Journal of Chongqing Jiaotong University (Natural Science), 2024, 43(10): 97-105.
[23]	Wali B, Khattak A J, Liu Jun. Heterogeneity assessment in incident duration modelling: implications for development of practical strategies for small & large scale incidents[J]. Journal of Intelligent Transportation Systems, 2022, 26(5): 586-601. doi: 10.1080/15472450.2021.1944135
[24]	Shang Qiang, Xie Tian, Yu Yang. Prediction of duration of traffic incidents by hybrid deep learning based on multi-source incomplete data[J]. International Journal of Environmental Research and Public Health, 2022, 19(17):DOI:10.3390/ijerph191710903.
[25]	罗文慧, 蔡凤田, 吴初娜, 等. 基于文本挖掘的道路运输安全风险源辨识模型[J]. 西南交通大学学报, 2021, 56(1):147-152.
	Luo Wenhui, Cai Fengtian, Wu Chu'na, et al. Text-mining based risk source identification model for transportation safety[J]. Journal of Southwest Jiaotong University, 2021, 56(1): 147-152.
[26]	Ji Keke, Li Zhengzhong, Chen Jian, et al. Freeway accident duration prediction based on social network information[J]. Neural Network World, 2022, 32(2): 93-112. doi: 10.14311/NNW.2022.32.006
[27]	Pereira C F, Rodrigues F, Ben-Akiva M. Text analysis in incident duration prediction[J]. Transportation Research Part C, 2013, 37:177-192. doi: 10.1016/j.trc.2013.10.002
[28]	Chen Jiaona, Tao Weijun. Traffic accident duration prediction using text mining and ensemble learning on expressways[J]. Scientific Reports, 2022, 12(1): DOI: 10.1038/s41598-022-25988-4.
[29]	周红磊, 张海涛, 栾宇, 等. 基于文本—图像增强的突发事件识别及分类方法研究[J]. 情报理论与实践, 2024, 47 (4): 181-188.
	Zhou Honglei, Zhang Haitao, Luan Yu, et al. Research on emergencies identification and classification method based on text-image enhancement[J]. Information Studies: Theory & Application, 2024, 47(4): 181-188.
[30]	纪柯柯, 陈坚, 肖思瑶, 等. 文本数据驱动下的高速公路事故持续时间预测模型[J]. 交通信息与安全, 2020, 38(6):9-16.
	Ji Keke, Chen Jian, Xiao Siyao, et al. A predictive model of highway accident duration driven by text data[J]. Journal of Transport Information and Safety, 2020, 38(6): 9-16.
[31]	陈娇娜, 陶伟俊, 靳引利, 等. 多模态文本信息的高速公路交通事件持续时间预测[J]. 中国安全生产科学技术, 2023, 19 (6): 180-186.
	Chen Jiaona, Tao Weijun, Jin Yinli, et al. Prediction of traffic incident duration on expressway based on multimodal text information[J]. Journal of Safety Science and Technology, 2023, 19(6): 180-186.
[32]	Chen Jiaona, Zhang Jin, Wang Peng, et al. A k-nearest text similarity-BiGRU approach for duration prediction of traffic accidents on expressways[J]. Discover Applied Sciences, 2025, 7: DOI:10.1007/s42452-025-07366-7.

定义		示例
事件	处置机构	施救队、消防
	处置人员	救援人员、养护人员
	处置设备	拖车、救护车、消防车
	处置措施	救援清理、强制分流、摆放锥形桶
	处置效果	火已扑灭、事故已处理完毕
	任务预判	预计1 h处理完毕
	任务难度	维修难度较大、救援难度较大
	事故持续时间	预计15 min、预计1 h
关系	属性关系	<编号,属性,货车>、 <编号,属性,侧翻>
	顺承关系	<路政,顺承,正在赶往现场>; <高交,顺承,临时管制>
	因果关系	<编号,因果,有路产损失>
	包含关系	<临时管制,包含,交警>
	并列关系	<灭火,并列,交通管制>
	对比关系	<预计15 min,对比,45 min>
实体	事故类型	剐蹭、追尾、侧翻、碰撞、自燃
	事故桩号	K+500、K+580
	事故地点类型	桥梁、主线、隧道、匝道
	事故车牌号	云D44^、京AFX^
	事故车辆类型	轿车、半挂、客车、危化品车、货车
	天气	晴、雾、阴、雨、雪
	路产损失	路产不详、有路产损失、无路产
	人员伤亡情况	无人员伤亡、2人轻伤
	交通管制	临时管制、临时封闭
属性	影响车道情况	占应急车道、占超车道
	实际持续时间	14 min、40 min
	所属路段	收费站至收费站
	事发时间	2021-01-01 T8:48:00
	交通分流	未分流、待核实、分流中
	编号	—

文献	MAPE值
[7]	2~15 min: 185.7%, >15 min: 45.4%,整体: 94.7%
[10]	5~15 min: 100.9%, 16~35 min: 75%~96%, 36~200 min: 20%~50%, >120 min: 50%~76%,整体: 27.58%~61%
[17]	异构集成模型: 35.610 1%
[26]	S-SVM: 21.7%, S-KNN: 20.8%
[28]	RF:75.155%
[32]	文本相似度BiGRU模型: 43.74%
文中	S-EEG:42.33%

模型	F
Word2vec-BiLSTM-CNN^[31]	0.520
S-EEG	0.532