Traffic accident severity prediction for secondary highways based on cluster analysis and SVM model

doi:10.16265/j.cnki.issn1003-3033.2022.05.1263

Abstract

Abstract:

In order to identify influence of input features on machine-learning-based traffic accident severity (TAS) prediction model, 12 potential factors were firstly selected as input variables based on 1808 accidents of a secondary mountainous highway. Then, KM algorithm was used to discretize continuous feature variables of TAS, and RF algorithm was adopted to select key feature variables. Finally, by combining three kinds of input feature variables (potential features, KM features, and RF features) and SVM algorithm, three kinds of TAS prediction models were developed respectively (SVM^*, KM-SVM, and RF-SVM), and their prediction performance was systematically analyzed in terms of accuracy and applicability. The results show that severity prediction accuracy of RF-SVM model is significantly improved by discretizing continuous variables and identifying key feature parameters, with the accuracy for severe injuries or deaths being improved about 40%. Influence of feature selection on SVM model performance is less than that of discretization of continuous variables. And RF-SVM model, in spite of a better prediction performance than binary logistic regression model, has higher sensitivity to different input features.

Key words: K-means (KM) clustering, support vector machine (SVM), accident severity, secondary mountainous highway, random forest (RF), machine learning

YANG Wenchen, ZHOU Yanning, TIAN Bijiang, GUO Fengxiang, HU Chengyu. Traffic accident severity prediction for secondary highways based on cluster analysis and SVM model[J]. China Safety Science Journal, 2022, 32(5): 163-169.

Figures/Tables 10

Fig.1

Fig.2

Tab.1

Definition of potential input feature variables

特征参数名称		值域		离散编码
肇事者性别 $x 1$		—		0=男,1=女
肇事者年龄 $x 2$ /岁		18~71		1=(<45),2=[45,60),3= (≥60)
事故发生季节 $x 3$		—		1=春季,2=夏季,3=秋季,4=冬季
事故发生时间 $x 4$		0 ~24		1=[0,6),2=[6,12),3=[12,18)4=[18,24)
R即x₅/m		>100		1=(<800),2=[800,1 600),3=(≥1 600)
特征参数名称	值域		离散编码
L $x 6$ /m	170~1800		0=(<1 100),1=(≥1 100)
D即 $x 7$ /(°)	0~6		1=(<2),2=[2,4),3= (≥4)
事故类型 $x 8$	—		1=追尾,2=碰撞固定物,3=侧翻,4=侧面碰撞,5=碰撞行人,6=正面碰撞,7=其他
肇事者车型 $x 9$	—		1=小型车,2=中型车,3=大型车,4=非机动车
涉事者车型 $x 10$	—		1=小型车,2=中型车,3=大型车,4=非机动车,5=其他
C_Q即 $x 11$	0.1-3.0		1=[0.00-0.65],2=(0.65-1.34],3=(1.34-2.00],4=(>2.00)
天气条件 $x 12$	—		0=晴天,1=非晴天

Tab.1

Fig.3

Fig.4

Tab.2

Fig.5

Fig.6

Fig.7

Fig.8

References 16

[1]	王建军, 曹旭东, 杨云峰. 基于CRSG模型的山区公路风险隔离研究[J]. 中国公路学报, 2018, 31(9):119-128.
	WANG Jianjun, CAO Xudong, YANG Yunfeng. Risk isolation analysis of mountain highways areas based on CRSG model[J]. China Journal of Highway and Transport, 2018, 31(9):119-128.
[2]	MANNERING F L, BHAT C R. Analytic methods in accident research: methodological frontier and future directions[J]. Analytic Methods in Accident Research, 2014, 1: 1-22. doi: 10.1016/j.amar.2013.09.001
[3]	MAMMERING F L, BHAT C, 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: DOI: 10.1016/j.amar.2020.100113. doi: 10.1016/j.amar.2020.100113
[4]	IRANITALAB A, KHATTAK A. Comparison of four statistical and machine learning methods for crash severity prediction[J]. Accident Analysis and Prevention, 2017, 108:27-36. doi: 10.1016/j.aap.2017.08.008
[5]	戢晓峰, 李德林, 杨文臣, 等. 山区二级公路交通事故致因的时间演化机制[J]. 中国安全科学学报, 2019, 29(4):31-36. doi: 10.16265/j.cnki.issn1003-3033.2019.04.006
	JI Xiaofeng, LI Delin, YANG Wenchen, et al. Temporal evolution analysis on causes of traffic accidents occurring on secondary highways in mountain areas[J]. China Safety Science Journal, 2019, 29(4):31-36. doi: 10.16265/j.cnki.issn1003-3033.2019.04.006
[6]	FOUNTAS G, ANASTASOPOULOS P C. A random thresholds random parameters hierarchical ordered probit analysis of highway accident injury-severities[J]. Analytic Methods in Accident Research, 2017, 15: 1-16. doi: 10.1016/j.amar.2017.03.002
[7]	杨文臣, 谢碧珊, 房锐, 等. 山区双车道公路机动车碰撞事故严重度致因比较分析与预测[J]. 交通运输系统工程与信息, 2021, 21(1):190-195.
	YANG Wenchen, XIE Bishan, FANG Rui, et al. Comparative analysis and prediction of motor vehicle crash severity on mountainous two-lane highways[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(1):190-195.
[8]	YU Rongjie, ABDEL-ATY M. Analyzing crash injury severity for a mountainous freeway incorporating real-time traffic and weather data[J]. Safety Science, 2014, 63(4):50-56. doi: 10.1016/j.ssci.2013.10.012
[9]	马壮林, 张祎祎, 杨杨, 等. 公路隧道交通事故严重程度预测模型研究[J]. 中国安全科学学报, 2015, 25(5):75-79.
	MA Zhuanglin, ZHANG Yiyi, YANG Yang, et al. Research on models for predicting severity of traffic accident in highway tunnel[J]. China Safety Science Journal, 2015, 25(5):75-79.
[10]	董傲然, 王长帅, 秦丹, 等. 机动车-行人事故中行人伤害严重程度分析[J]. 中国安全科学学报, 2020, 30(11): 141-147. doi: 10.16265/j.cnki.issn 1003-3033.2020.11.021
	DONG Aoran, WANG Changshuai, QIN Dan, et al. Analysis on injury severity of pedestrian in motor vehicle-pedestrian accidents[J]. China Safety Science Journal, 2020, 30(11): 141-147. doi: 10.16265/j.cnki.issn 1003-3033.2020.11.021
[11]	ZHU Li, YU Fei Richard, WANG Yige, et al. Big data analytics in intelligent transportation systems: a survey[J]. IEEE Transactions on Intelligent Transportation Systems, 2018, 20(1): 383-398. doi: 10.1109/TITS.2018.2815678
[12]	YU Rongjie, ABEDEL-ATY M. Utilizing support vector machine in real-time crash risk evaluation[J]. Accident Analysis and Prevention, 2013, 51:252-259. doi: 10.1016/j.aap.2012.11.027
[13]	游锦明, 王俊骅, 唐棠, 等. 基于支持向量机的高速公路实时事故风险研判[J]. 同济大学学报:自然科学版, 2017, 45(3):355-361.
	YOU Jinming, WANG Junye, TANG Tang, et al. Support vector machines approach for predicting real-time rear-end crash risk on freeways[J]. Journal of Tongji University: Natural Science, 2017, 45(3):355-361.
[14]	SUN Ming, SUN Xiaoduan, SHAN Donghui. Pedestrian crash analysis with latent class clustering method[J]. Accident Analysis and Prevention, 2019, 124:50-57. doi: 10.1016/j.aap.2018.12.016
[15]	SILVA P B, ANDRADE M, FERREIR A S. Machine learning applied to road safety modeling: a systematic literature review[J]. Journal of Traffic and Transportation Engineering: English edition, 2020, 7(6):775-790.
[16]	周志华. 机器学习[M]. 北京: 清华大学出版社, 2016: 121-204.
	ZHOU Zhihua. Machine learning[M]. Beijing: Tsinghua University Press, 2016:121-204.

模型	C	γ	预测准确率
模型	C	γ	非严重事故/%	严重事故/%	整体/ %
SVM^*	9	0.07	91.15	18.38	64.09
KM-SVM	1.1	0.09	83.63	58.82	81.77
RF-SVM	32	0.06	88.93	67.65	82.60