Research on mediating effect of express way accident duration based on text information

doi:10.16265/j.cnki.issn1003-3033.2023.04.0826

Abstract

Abstract:

In order to explore the internal coupling relationship of the influencing factors of traffic accident duration and reveal the mechanism of text information in the chain transmission process, a keyword importance analysis method based on word frequency-inverse text frequency (TF-IDF) model was proposed by using natural language processing technology and random forest algorithm. At the same time, a multiple mediating effect model of highway traffic accident duration is established, and the Bootstrap sampling method of product coefficient was used to test the mediating effect. The significance of text features in parallel mediating path and two-level chain mediating path was tested respectively, and the strength of mediating effect was calculated. Taking 3 046 traffic accident records of expressways in Shanxi Province as an example, the results showed that in the influence path of duration, there is a partial chain mediating relationship between accident type and month, and the mediating effect accounts for 11.868%. There is a complete chain mediating relationship between accident type and weather, location and accident range, and the mediating effect accounts for 100%. The number of characters and the number of reports are significant mediating variables in the specific path. The number of reports is completely mediated by the type of mediating effect of the weather on the duration. The number of characters is completely mediated by the type of mediating effect of the accident range on the duration. The impact of the period on the duration is 7.075% through the number of characters. The two-level chain mediating path of specific keyword sets exists.

Key words: text information, express way, duration of accident, mediating effect, traffic accidents, random forest

CHEN Jiaona, JIN Yinli, TAO Weijun, LI Daofeng. Research on mediating effect of express way accident duration based on text information[J]. China Safety Science Journal, 2023, 33(4): 155-162.

Figures/Tables 11

Fig.1

Tab.1

Tab.2

Fig.2

Fig.3

Fig.4

Fig.5

Tab.3

Tab.4

Tab.5

Tab.6

References 20

[1]	ADLER M W, OMMEREN J V, RIETVELD P. Road congestion and incident duration[J]. Economics of Transportation. 2013, 2(4):109-118. doi: 10.1016/j.ecotra.2013.12.003
[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):10-22. doi: 10.1007/s12544-017-0283-3
[3]	MA Zhengjing, MEI Gang, CUOMO S. An analytic framework using deep learning for prediction of traffic accident injury severity based on contributing factors[J]. Accident Analysis & Prevention. 2021, 160(9):DOI: 10.1016/J.AAP.2021.106322. doi: 10.1016/J.AAP.2021.106322
[4]	杨文臣, 张轮, 施奕骋, 等. 城市快速路交通事件持续时间生存分析[J]. 交通运输系统工程与信息, 2014, 14(5):168-174.
	YANG Wenchen, ZHANG Lun, SHI Yicheng. Survival analysis of traffic incident duration for urban expressways[J]. Journal of Transportation Systems Engineering and Information Technology, 2014, 14(5):168-174.
[5]	DING Chuan, MA Xiaolei, WANG Yinhai, et al. Exploring the influential factors in incident clearance time: disentangling causation from self-selection bias[J]. Accident Analysis and Prevention, 2015, 85:58-65. doi: 10.1016/j.aap.2015.08.024
[6]	郑先俊, 尹静, 陈瑞华, 等. 青藏高速公路危险物品运输环境风险评价及应急对策[J]. 水土保持通报, 2019, 39(3):270-275.
	ZHENG Xianjun, YIN Jing, CHEN Ruihua, et al. Environment risk assessment and emergency countermeasures of dangerous goods transportating on Qinghai-Tibet expressway[J]. Bulletin of Soil and Water Conservation, 2019, 39(3):270-275.
[7]	苏磊. 基于深度数据挖掘的危险品运输事故关键影响因素及致因模型研究[D]. 青岛: 青岛科技大学, 2019.
	SU Lei. Key influencing factors and cause model of dangerous goods transportation accidents analysisi based on deep data mining[D]. Qingdao: Qingdao University of Science and Technology, 2019.
[8]	ZHANG Zihe, LIU Jun, LI Xiaobing, et al. Do larger sample sizes increase the reliability of traffic incident duration models? a case study of east tennessee incidents[J]. Transportation Research Record, 2021, 2675(6): 265-280.
[9]	MOHAMMED Z A, ABDULLAH M N, Al-HUSSAINI I H. Review of the traffic incident duration prediction methods[J]. Journal of Research in Science and Engineering, 2020, 2(6): 20-25.
[10]	ABDULSAMET S, HALIT O. Estimation of traffic incident duration: a comparative study of decision tree models[J]. Arabian Journal for Science and Engineering, 2020, 45: 1-12. doi: 10.1007/s13369-019-04073-5
[11]	DISERENS R V, MARMY C, PASQUIER M, et al. Modelling transport time to trauma centres and 30-day mortality in road accidents in Switzerland: an exploratory study[J]. Swiss Medical Weekly, 2021, 151:35-36.
[12]	LI Xiaobing, LIU Jun, KHATTAK A, et al. Sequential prediction for large-scale traffic incident duration: application and comparison of survival models[J]. Transportation Research Record, 2020, 2674(1):79-93.
[13]	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, 2019:1-11.
[14]	FRANCISCO C P, FILIPE R, MOSHE B A. Text analysis in incident duration prediction[J]. Transportation Research Part C, 2013, 37:177-192. doi: 10.1016/j.trc.2013.10.002
[15]	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]. Traffic Information and Safety, 2020, 38(6):9-16.
[16]	牟庆泉, 丁小兵, 刘志钢, 等. 基于地铁运营日志文本挖掘的危险源辨识算法研究[J]. 中国安全生产科学技术, 2022, 18(3):204-210.
	MU Qingquan, DING Xiaobing, LIU Zhigang, et al. Research on identification algorithm of hazard sources based on text mining of metro operation log[J]. Journal of Safety Science and Technology, 2022, 18(3):204-210.
[17]	韩天园, 田顺, 吕凯光, 等. 基于文本挖掘的重特大交通事故成因网络分析[J]. 中国安全科学学报, 2021, 31(9):150-156. doi: 10.16265/j.cnki.issn1003-3033.2021.09.021
	HAN Tianyuan, TIAN Shun, LYU Kaiguang, et al. Network analysis on causes for serious traffic accidents based on text mining[J]. China Safety Science Journal, 2021, 31(9):150-156. doi: 10.16265/j.cnki.issn1003-3033.2021.09.021
[18]	于卫红, 付飘云, 任月, 等. 基于PMI与BTM的船舶事故原因文本挖掘[J]. 交通信息与安全, 2021, 39(1):35-44.
	YU Weihong, FU Piaoyun, REN Yue, et al. Text mining for causes of ship accidents based on PMI and BTM[J]. Traffic Information nd Safety, 2021, 39(1):35-44.
[19]	孙欢. 基于BERT+BiLSTM+CRF模型和改进Apriori算法的交通事故文本分析[D]. 西安: 长安大学, 2021.
	SUN Huan. Traffic accident text analysis based on BERT+Bi LSTM+CRF model and improved apriori algorithm[D]. Xi'an: Chang'an University, 2021.
[20]	屠君超. 基于深度学习的交通文本检测和识别方法的研究[D]. 成都: 电子科技大学, 2021.
	TU Junchao. Research on methods of traffic text detection and recognition based on deep learning[D]. Chengdu: University of Electronic Science and Technology, 2021.

a_ij和b_j	c_ij'或c_ijk'	a_ij×b_j与c_ij'或 a_ij×b×b'_jk与c_ijk'	BootCI	检验结果
均显著	不显著	—	—	完全中介作用
均显著	显著	同号	—	部分中介作用
均显著	显著	异号	—	遮掩作用
至少一个不显著	—	—	包含0	中介作用不显著
至少一个不显著	不显著	—	不包含0	完全中介作用
至少一个不显著	显著	同号	不包含0	部分中介作用
至少一个不显著	显著	异号	不包含0	遮掩作用

变量类别	变量	说明
因变量Y	持续时间Y	报警时刻与处置完成时刻之差
自变量X	月份X₁	事件报警时刻所属的月份,1—12
	时段X₂	事件报警时刻所属的时段,0∶00—23∶00
	星期X₃	1星期一;2星期二;3星期三;4星期四;5星期五;6星期六;7星期日
	事故类型X₄	1单方侧翻;2单方碰撞;3单方自燃;4单方故障;5多方追尾;6多方相撞;7多方剐蹭
	事故范围X₅	开始桩号与结束桩号之差
	天气X₆	1晴;2阴;3雨;4雪;5雾
	位置类型X₇	1道路;2主线;3收费站; 4桥梁;5隧道
控制变量C	受伤人数C₁	—
	死亡人数C₂	—
	损毁车辆数C₃	—
	危化品车辆C₄	是否涉及危化品车辆:1涉及;0不涉及
中介变量M(待检验)	上报次数M₁	事件持续期间信息报送次数
	字符数M₂	事件发现时首次文本信息报送字符数
	情报板发布M₃	事件发现时是否在全线情报板进行发布:0否,1是
	主题关键词M_j	—

路径	检验结论	效应占比/%
月份→事故类型→持续时间	部分中介	11.868
时段→事故类型→持续时间	不显著	0
星期→事故类型→持续时间	不显著	0
天气→事故类型→持续时间	完全中介	100
位置→事故类型→持续时间	完全中介	100
事故范围→事故类型→持续时间	完全中介	100

路径	检验结论	效应占比/%
月份→字符数→持续时间	遮掩效应	7.296
月份→上报次数→持续时间	不显著	0
月份→情报板发布→持续时间	不显著	0
时段→字符数→持续时间	部分中介	7.075
时段→上报次数→持续时间	不显著	0
时段→情报板发布→持续时间	不显著	0
星期→字符数→持续时间	不显著	0
星期→上报次数→持续时间	不显著	0
星期→情报板发布→持续时间	不显著	0
天气→字符数→持续时间	不显著	0
天气→上报次数→持续时间	完全中介	100
天气→情报板发布→持续时间	不显著	0
位置→字符数→持续时间	不显著	0
位置→上报次数→持续时间	不显著	0
位置→情报板发布→持续时间	不显著	0
事故范围→字符数→持续时间	完全中介	100
事故范围→上报次数→持续时间	不显著	0

链式路径	区间是否包含0	检验结果
月份→事故类型→关键词集→ 持续时间	否	存在
时段→事故类型→关键词集→ 持续时间	是	不存在
星期→事故类型→关键词集→ 持续时间	是	不存在
事故范围→事故类型→关键词集→ 持续时间	否	存在
天气→事故类型→关键词集→ 持续时间	否	存在
位置→事故类型→关键词集→ 持续时间	否	存在