| 119 | 0 | 34 |
| Downloads | Citas | Reads |
Using real-world operational data from an expressway in Shandong Province, China, this study developed a framework for short-term traffic crash risk prediction aimed at routine safety management. Traffic flow, meteorological conditions, and crash records from multiple monitoring sites along the expressway were aggregated into 5-minute intervals and spatiotemporally fused with crash occurrences. Candidate explanatory variables were constructed to characterize upstream and downstream traffic volume, speed, occupancy, visibility, rainfall, temperature, wind, time of day, and other temporal indicators. Recursive Feature Elimination(RFE) was employed to rank variable importance, resulting in the retention of the 15 most informative features within the pre-crash window, thereby highlighting the joint influence of traffic conditions and adverse weather. Within an ensemble learning framework, an expressway crash risk predictor was developed by integrating eXtreme Gradient Boosting(XGBoost) with a Convolutional Neural Network(CNN). XGBoost captures nonlinear relationships among heterogeneous tabular features, while the CNN extracts local temporal patterns from traffic flow sequences. Their probabilistic outputs were combined through an optimized weighting scheme, and model hyperparameters were tuned via cross-validation. For fair comparison, baseline models using only XGBoost and only CNN were trained on the same dataset following identical preprocessing and evaluation protocols. Experimental results indicate that the proposed integrated model achieves an accuracy of 0.96 and an AUC of 0.94 on the test set, outperforming both single models in terms of discrimination and robustness. Feature ablation analyses reveal that excluding meteorological variables results in the greatest performance degradation, with visibility and rainfall identified as the most influential factors. Prediction errors increase significantly under rainy or low-visibility conditions, confirming the critical role of meteorological factors in crash risk prediction. Furthermore, testing on a later time period of the same roadway yielded an accuracy of 87.4%, demonstrating effective out-of-time generalization. Overall, the XGBoost-CNN ensemble provides an accurate, robust, and interpretable tool for estimating short-term crash risk on expressways.
[1] 中华人民共和国统计局.中国统计年鉴[M].北京：中国统计出版社，2024.National Bureau of Statistics of the People's Republic of China.China statistical yearbook[M].Beijing:China Statistics Press,2024.
[2] 张驰，杨绍祥，冯逸伟，等.基于风险更新理论的高速公路动态风险判别方法[J].长安大学学报(自然科学版),2021,41(4):115-126.Zhang Chi,Yang Shaoxiang,Feng Yiwei,et al.Expressway dynamic risk discrimination method based on risk update theory[J].Journal of Chang'an University (Natural Science Edition),2021,41(4):115-126.
[3] 王博，刘昌赫，张驰，等.基于道路监控的高速公路作业区碰撞风险预警[J].浙江大学学报(工学版),2024,58(6):1221-1232.Wang Bo,Liu Changhe,Zhang Chi,et al.Crash risk early warning in highway work zone based on road surveillance camera[J].Journal of Zhejiang University (Engineering Science),2024,58(6):1221-1232.
[4] Cheng Rui,Dai Junji,Mi Yunjian,et al.Research on the prediction model of high-risk accident points in mountainous road traffic environments[J].Transportation Safety and Environment,2025,7(3):tdaf045.
[5] 孟祥海，由冰玉，邱志雄，等.基于事故修正系数的山区高速公路事故预测模型[J].中国安全科学学报，2023,33(1):10-17.Meng Xianghai,You Bingyu,Qiu Zhixiong,et al.Accident prediction model of mountainous freeway based on crash modification factors[J].China Safety Science Journal.2023,33(1):10-17.
[6] Mannering F L,Shankar V,Bhat C R.Unobserved heterogeneity and the statistical analysis of highway accident data[J].Analytic Methods in Accident Research,2016,11:1-16.
[7] Abdel-Aty M,Uddin N,Pande A,et al.Predicting freeway crashes from loop detector data by matched case-control logistic regression[J].Transportation Research Record,2004,1897(1):88-95.
[8] Buddhavarapu P,Scott J G,Prozzi J A.Modeling unobserved heterogeneity using finite mixture random parameters for spatially correlated discrete count data[J].Transportation Research Part B:Methodological,2016,91:492-510.
[9] Mannering F,Bhat C R,Shankar V,et al.Big data,traditional data and the tradeoffs between prediction and causality in highway-safety analysis[J].Analytic Methods in Accident Research,2020,25:100113.
[10] Hossain M,Muromachi Y.Understanding crash mechanisms and selecting interventions to mitigate real-time hazards on urban expressways [J].Transportation Research Record,2011,2213(1):53-62.
[11] Hossain M,Muromachi Y.A Bayesian network based framework for real-time crash prediction on the basic freeway segments of urban expressways[J].Crash Analysis & Prevention,2012,45:373-381.
[12] 谢学斌，孔令燕.基于ARIMA和XGBoost组合模型的交通事故预测[J].安全与环境学报，2021,21(1):277-284.Xie Xuebin,Kong Lingyan.On the ways to the traffic accident prediction based on the ARIMA and XGBoost combined model[J].Journal of Safety and Environment,2021,21(1):277-284.
[13] 游锦明，方守恩，张兰芳，等.高速公路实时事故风险研判模型及可移植性[J].同济大学学报(自然科学版),2019,47(3):347-352.You Jinming,Fang Shouen,Zhang Lanfang,et al.Real-Time crash risk prediction models and transferability analysis on freeways[J].Journal of Tongji University (Natural Science),2019,47(3):347-352.
[14] 管林涛，黄志强，陈洋.一种基于时空Transformer的交通事故预测方法[J].电脑与信息技术，2022,30(1):8-13.Guan Lintao,Huang Zhiqiang,Chen Yang.A method of traffic accident prediction based on spatio-temporal transformer[J].Computer and Information Technology,2022,30(1):8-13.
[15] Amin M S R,Zareie A,Amador-Jiménez L E.Climate change modeling and the weather-related road crashes in Canada[J].Transportation Research Part D:Transport and Environment,2014,32:171-183.
[16] 黄冰娥.恶劣天气下高速公路行车风险分析与预警管理研究[D].上海：上海交通大学，2012.Huang Binge.Study on the traffic risk analysis and forewarning management of freeway under disaster weather[D].Shanghai:Shanghai Jiao Tong University,2012.
[17] 戴学臻，胡传喜.基于改进DEMATEL-ISM的高速公路重特大交通事故影响因素研究[J].安全与环境学报，2022,22(3):1217-1226.Dai Xuezhen,Hu Chuanxi.Research on influencing factors of serious traffic accidents on freeway based on improved DEMATEL-ISM[J].Journal of Safety and Environment,2022,22(3):1217-1226.
[18] Ye Fei,Cheng Wen,Wang Changshuai,et al.Investigating the severity of expressway crash based on the random parameter logit model accounting for unobserved heterogeneity[J].Advances in Mechanical Engineering,2021,13(12):16878140211067278.
[19] 徐铖铖，刘攀，王炜，等.恶劣天气下高速公路实时事故风险预测模型[J].吉林大学学报(工学版),2013,43(1):68-73.Xu Chengcheng,Liu Pan,Wang Yi,et al.Real time crash risk prediction model on freeways under nasty weather conditions[J].Journal of Jilin University (Engineering and Technology Edition),2013,43(1):68-73.
[20] Joo Y J,Kho S Y,Kim D K,et al.A data-driven Bayesian network for probabilistic crash risk assessment of individual driver with traffic violation and crash records[J].Crash Analysis and Prevention,2022,176:106790.
[21] Chen Tianqi,Guestrin C.XGBoost:a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York:Association for Computing Machinery,2016:785-794.
[22] Alzubaidi L,Zhang J,Humaidi A J,et al.Review of deep learning:concepts,CNN architectures,challenges,applications,future directions[J].Journal of Big Data,2021,8:1-74.
[23] 何红艳，黄国言，张炳，等.基于极限树特征递归消除和LightGBM的异常检测模型[J].信息网络安全，2022,22(1):64-71.He Hongyan,Huang Guoyan,Zhang Bing,et al.Intrusion detection model based on extra trees-recursive feature elimination and LightGBM[J].Netinfo Security,2022,22(1):64-71.
[24] 杨春霞，秦家鹏，王庆，等.基于多车道加权融合的短时交通流预测研究[J].公路交通科技，2021,38(1):121-127.Yang Chuanxia,Qin Jiapeng,Wang Qing,et al.Study on short-term traffic flow prediction based on multi-lane weighted fusion[J].Journal of Highway and Transportation Research and Development,2021,38(1):121-127.
[25] 樊博.基于实时交通流的高速公路事故风险预测与致因分析方法研究[D].重庆：重庆交通大学，2021.Fan Bo.The study of the crash risk prediction and crash causation analysis for freeways based on the real-time traffic flow[D].Chongqing:Chongqing Jiaotong University,2021.
[26] Li Pei,Abdel-Aty M,Yuan Jinghui.Real-time crash risk prediction on arterials based on LSTM-CNN[J].Accident Analysis & Prevention,2020,135:105371.
[27] 陈云，韦天焕.欠采样与过采样技术研究[J].信息通信，2015(4):18-19.Chen Yun,Wei Tianhuan.Research on undersampling and oversampling techniques[J].Changjiang Information & Communications,2015(4):18-19.
[28] 谢辉.基于采样技术和机器学习算法的交通事故严重程度预测[J].公路交通科技(应用技术版),2020,16(2):312-314.Xie Hui.Traffic accident severity prediction based on sampling technology and machine learning algorithm[J].Highway Traffic Technology (Applied Technology Edition),2020,16(2):312-314.
[29] Yao Chang,Wu Feng,Chen Houjin,et al.Traffic sign recognition using HOG-SVM and grid search[C]//Proceedings of International Conference on Signal Processing.Piscataway,NJ,USA:Institute of Electrical and Electronics Engineers,2015.
[30] 温惠英，黄坤火，赵胜.基于机器学习的高速公路大型货车追尾风险预测[J].中国安全科学学报，2023,33(9):173-180.Wen Huiying,Huang Kunhuo,Zhao Sheng.Prediction of rear-end collision risk of freeway trucks based on machine learning[J].China Safety Science Journal,2023,33(9):173-180.
[31] 张开放，苏华友，窦勇.一种基于混淆矩阵的多分类任务准确率评估新方法[J].计算机工程与科学，2021,43(11):1910-1919.Zhang Kaifang,Su Huayou,Dou Yong.A new multi-classification task accuracy evaluation method based on confusion matrix[J].Computer Engineering & Science,2021,43(11):1910-1919.
[32] 刘钦，宋太龙，李振龙，等.小样本下基于迁移学习与LSTM的雾天高速公路车辆跟驰模型[J].交通信息与安全，2023,41(1):13-22.Liu Qin,Song Tailong,Li Zhenlong,et al.A car-following model for expressway under foggy weather based on transfer learning and LSTM with small-sample[J].Journal of Transport Information and Safety,2023,41(1):13-22.
Basic Information:
DOI：10.13637/j.issn.1009-6094.2025.1216
China Classification Code:U491.31
Citation Information:
[1]Liu Wenke,Wang Xuekai,Zhou Mo ,et al.Expressway traffic crash risk prediction based on an integrated approach of XGBoost-CNN[J].Journal of Safety and Environment,2026,26(08):3172-3181.DOI:10.13637/j.issn.1009-6094.2025.1216.
Fund Information:
国家自然科学基金项目(52362050); 陕西省博士后科研项目(2023BSHEDZZ216); 秦创原引用高层次创新创业人才项目(QCYRCXM2023-110); 中国博士后科学基金项目(2024M752739)
2026-08-19
2026-08-19