Civil Engineering |
|
|
|
|
Traffic incident detection based on variable selection and kernel extreme learning machine |
SHANG Qiang1, LIN Ci-yun1,2, YANG Zhao-sheng1,2, BING Qi-chun1,3, XING Ru-ru1 |
1. College of Transportation, Jilin University, Changchun 130022, China;
2. Jilin Province Key Laboratory of Road Traffic, Jilin University, Changchun 130022, China;
3. College of Automobile and Transportation, Qingdao Technological University, Qingdao 266520, China |
|
|
Abstract An automatic incident detection (AID) algorithm was proposed based on variable selection and kernel extreme learning machine (KELM) in order to further improve the performance of traffic incident detection. A relatively comprehensive initial variable set was constructed for traffic incident detection according to the changing characteristics of upstream and downstream traffic parameters in a traffic incident, and important variables were selected from the initial variable set by the random forest-recursive feature elimination (RF-RFE) algorithm. Then the important variables were used as input of KELM and the KELM was trained. The parameters were optimized by the gravitational search algorithm (GSA). The I-880 database of the United States was used to validate and comparatively analyze the performance of the proposed AID algorithm. Because the number of incident samples is much less than the number of incident-free samples in the database, the SMOTE is used to balance the two kinds of samples. Results show that using important variables can improve the performance of traffic incident detection, and the performance of KELM is better than that of the back propagation neural network (BPNN) and support vector machine (SVM).
|
Received: 02 June 2016
Published: 08 July 2017
|
|
基于变量选择和核极限学习机的交通事件检测
为了进一步提高交通事件检测的效果,提出基于变量选择和核极限学习机(KELM)的自动事件检测(AID)算法.根据交通事件上、下游交通流参数的变化特点,构建较全面的交通事件检测初始变量集.采用随机森林—递归特征消除(RF-RFE)算法,从中选择重要变量.以重要变量作为输入,训练KELM并通过万有引力搜索算法(GSA)优化参数.使用美国I-880数据库,对AID算法的效果进行验证和对比分析.因为数据库中的事件样本数远少于非事件样本数,采用SMOTE平衡两类样本.结果表明,使用重要变量能够提高交通事件的检测效果,KELM的检测效果优于反向传播神经网络(BPNN)和支持向量机(SVM).
|
|
[1] LINDLEY J A. Urban freeway congestion: quantification of the problem and effectiveness of potential solutions [J]. ITE Journal, 1987, 57(1): 27-32.
[2] 姬杨蓓蓓,张小宁,孙立军. 基于贝叶斯决策树的交通事件持续时间预测[J].同济大学学报:自然科学版,2008,36(3): 319-324. JIYANG Bei-bei, ZHANG Xiao-ning, SUN Li-jun.Incident duration prediction grounded on Bayesian decision method based tree algorithm [J]. Journal of Tongji University: Natural Science, 2008, 36(3): 319-324.
[3] GHOSH B, SMITH D P. Customization of automatic incident detection algorithms for signalized urban arterials [J]. Journal of Intelligent Transportation Systems, 2014, 18(4): 426-441.
[4] KARIM A, ADELI H. Comparison of fuzzy-waveletradial basis function neural network freeway incidentdetection model with California algorithm [J]. Journal of Transportation Engineering, 2002, 128(1): 21-30.
[5] SRINIVASAN D, JIN X, CHEU R L. Evaluation ofadaptive neural network models for freeway incidentdetection [J]. IEEE Transactions on Intelligent Transportation Systems, 2004, 5(1): 1-11.
[6] 陈维荣,关佩,邹月娴.基于SVM的交通事件检测技术[J].西南交通大学学报,2011,46(1): 63-67. CHEN Wei-rong, GUAN Pei, ZOU Yue-xian. Automatic incident detection technology based on SVM [J]. Jounal of Southwest Jiaotong University, 2011, 46(1): 63-67.
[7] YUAN F, CHEU R L. Incident detection using support vector machines [J]. Transportation Research Part C:Emerging Technologies, 2003, 11(3): 309-328.
[8] CHEN S, WANG W. Decision tree learning for freeway automatic incident detection [J]. Expert Systems with Applications, 2009, 36(2): 4101-4105.
[9] WANG J, LI X, LIAO S S, et al. A hybrid approach for automatic incident detection [J]. IEEE Transactions on Intelligent Transportation Systems, 2013, 14(3): 1176-1185.)
[10] XIAO J, GAO X, KONG Q J, et al. More robust and better: a multiple kernel support vector machine ensemble approach for traffic incident detection [J]. Journal of Advanced Transportation, 2014, 48(7):858-875.
[11] LIU Q, LU J, CHEN S, et al. Multiple Naïve bayes classifiers ensemble for traffic incident detection [J]. Mathematical Problems in Engineering, 2014, 2014: 1-16.
[12] HUANG G B, ZHU Q Y, SIEW C K. Extreme learning machine: theory and applications [J]. Neurocomputing, 2006, 70(1): 489-501.
[13] HUANG G B, ZHOU H, DING X, et al. Extreme learning machine for regression and multiclass classification [J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 2012, 42(2):513-529.
[14] PAL M, MAXWELL A E, WARNER T A. Kernel-based extreme learning machine for remote-sensing image classification [J]. Remote Sensing Letters, 2013,4(9): 853-862.
[15] XIE W, LI Y, MA Y. Breast mass classification in digital mammography based on extreme learning machine [J]. Neurocomputing, 2016, 173(3): 930-941.
[16] GRANITTO P M, FURLANELLO C, BIASIOLI F, et al. Recursive feature elimination with random forest for PTR-MS analysis of agroindustrial products [J]. Chemometrics and Intelligent Laboratory Systems, 2006, 83(2): 83-90.
[17] RASHEDI E, NEZAMABADI-POUR H, SARYAZDI S. GSA: a gravitational search algorithm [J]. Information sciences, 2009, 179(13): 2232-2248.
[18] 楼晓俊,孙雨轩,刘海涛.聚类边界过采样不平衡数据分类方法[J].浙江大学学报:工学版, 2013, 47(6): 944-950. LOU Xiao-jun, SUN Yu-xuan, LIU Hai-tao. Clustering boundary over-sampling classification method for imbalanced data sets [J]. Journal of Zhejiang University: Engineering Science, 2013, 47(6): 944-950.
[19] BREIMAN L. Random forests [J]. Machine Learning, 2001, 45(1): 5-32.
[20] GUYON I, WESTON J, BARNHILL S, et al. Gene selection for cancer classification using support vector machines [J]. Machine Learning, 2002, 46(1/3):389-422.
[21] TENG H, QI Y. Application of wavelet technique to freeway incident detection [J]. Transportation Research Part C: Emerging Technologies, 2003, 11(3):289-308. |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
|
Shared |
|
|
|
|
|
Discussed |
|
|
|
|