基于贝叶斯网络的交通事故态势研究
|
摘要
随着我国社会经济的快速发展，机动化程度不断提高，道路通车里程和车辆运行里程显著增加，客货出行频率和道路交通量明显增大，导致道路交通事故频发，道路交通安全态势仍然严峻。交通事故不但导致人员死伤和直接财产损失，还将导致交通堵塞、交通延误、火灾、爆炸、危险品泄漏等问题，给人民的生产生活和社会经济发展带来巨大损失。为了保障社会生产生活的顺利进行，减少交通事故造成的人员死伤和财产损失，需要加大力度防控道路交通事故，加强道路交通安全。
道路交通事故态势分析的目的是探究道路交通事故与人、车、路、环境等系统之间的相互关系，掌握在各项因素影响下交通事故的发生、发展和生成最终结果的内在规律，以期提出有效的交通安全管理措施，做到在事故发生前预防事故的发生，在事故发生过程中降低事故的严重程度，在事故发生后迅速进行事故快速响应，减少人员伤亡和财产损失。可见，道路交通事故态势研究对于提出并实施行之有效的交通安全管理对策，提高我国的道路交通安全水平，降低交通事故的危害程度，保障社会生产和人民生活，具有非常重要的作用。
本文基于这些背景，结合国家863计划项目《重特大交通事故快速勘测与处置技术》，根据我国的交通事故状况、交通系统特性和社会经济发展特性，建立交通事故态势分析和预测模型，分析各种影响因素与交通事故之间的错综复杂的因果关系，掌握交通事故的发生和发展变化规律，并将研究成果应用于交通安全管理措施的制定和评价以及交通事故快速响应方案的制定，以提高我国的道路交通安全水平，促进道路交通和社会经济的快速发展。
本文完成的主要研究工作有：
论述了交通事故态势分析的基本理论，研究交通事故态势分析的两种主要方法——贝叶斯网络和离散选择模型，并分别应用两种方法建立机动车交通事故的态势分析模型，根据模型检验结果和预测精度进行模型优选；
应用贝叶斯理论，建立非机动车/行人交通事故态势分析模型；
基于已建的机动车和非机动车/行人交通事故态势分析模型，应用团树传播算法，学习各项影响因素对交通事故态势的影响规律；
应用生存分析理论，建立交通事故持续时间预测模型；
根据交通事故态势分析结果，提出降低交通事故严重程度的安全管理措施；
制定交通事故快速响应方案，将已建的交通事故态势分析模型和交通事故持续时间预测模型整合，建立交通事故快速响应系统框架，并研究交通事故态势预测的贝叶斯网络增量学习问题。
本文在以下方面取得了一定创新成果：
构建机动车交通事故和非机动车/行人交通事故态势分析的贝叶斯网络模型，应用贝叶斯网络的结构和参数，对自变量与因变量之间、多维因变量之间以及多个自变量之间的复杂影响关系进行结构化描述和量化表达；
以变量选择的客观性、变量间影响关系的表达、建模精确度等多指标为依据，从定性分析和定量计算两方面对比分析贝叶斯网络模型和离散选择模型在交通事故态势分析方面的优劣；
应用团树传播算法实现贝叶斯网络的推理学习，分析各影响因素的变化对事故严重程度的影响，并将其用于交通事故安全对策的制定和定量评价；
为了解决贝叶斯网络的增量学习问题，提出以新增数据比例10%作为贝叶斯网络结构调整的阈值；
应用生存分析理论构建交通事故持续时间预测的加速失效模型。
本研究成果可用于揭示在各项因素影响下交通事故的发生、发展和生成最终结果的内在规律，提出系统性的交通安全管理措施，预测交通事故态势和持续时间，制定有针对性的交通事故快速响应方案。研究成果对提高我国的道路交通安全水平，降低事故导致的人员伤亡和财产损失，保障社会生产和人民生活，具有重要的理论价值和现实意义。
With the huge progress of China in social, economic, motorization andtransportation recently, the level of road traffic safety is getting lower. Trafficaccidents cause not only casualty and property damage, but also traffic delay and jam,fire, explosion, and leakage of dangerous goods sometimes, which lead to huge lossfor the people. In order to make the social and economic develop rapidly and reducethe damage caused by traffic accident, it is necessary for the government to figure outhow to enhance traffic safety and prevent from traffic accident.
The goal of situation analysis is to examine the relationship among trafficaccident and the potential factors, including people, vehicle, road and environment, aswell as to analyze the process of developing and changing of traffic accident. Basedon situation analysis, effective traffic safety management measures can be made sothat the damage caused by traffic accident can be reduced. Therefore, the researchconcerning about situation analysis is of great importance for enhancing traffic safetyand ensuring the rapid social and economic development.
Founded by National High-tech R&D Program of China (863Program)-QuickInvestigation and Disposal Technologies of Major Accidents, a forecasting model wasdeveloped for traffic accident situation, and the relationship between factors andtraffic accident was analyzed according to the conditions and features of the trafficsystem and social-economic development in China. It then proposed some trafficmanagement measures and developed a program of traffic accident rapid response toimprove traffic safety.
Two potential methods were studied for situation analysis, which are Bayesiannetwork and discrete choice model, and employed them respectively to modelsituation of motor vehicle accident. Bayesian network, which was proved to be betterthan discrete choice model in prediction accuracy, was picked out and employed inmodeling the situations of not only motor vehicle accident, but also non-motorvehicle/pedestrian accident. Then the impacts of the factors on accident situation arelearned by using the Clique Tree Propagation based on the Bayesian networks, andsome traffic management strategies were proposed to reduce damage caused by trafficaccident. A forecasting model of accident duration was also developed according to the theory of survival analysis. Based on the models of situation analysis and durationforecasting, a rapid response system was then developed. The incremental learning ofBayesian network was concerned about at the end of the dissertation.
The major contributions of this dissertation are as follows:
(1) developing situation analysis models for motor vehicle accident andnon-motor vehicle/pedestrian accident, respectively, with Bayesian network, andanalyzing the relationship among dependent variables and independent variables byusing the methods of both structural description and quantitative deduction;
(2) comparing Bayesian network and discrete choice model with respect to thefactors of both qualitative analysis and quantitative calculation, including variableschoosing, relationship analysis and goodness-of-fit of the models, etc.;
(3) inferring the impacts of factors on accident situation with the Clique TreePropagation, and proposing and evaluating traffic management measures according tothe inferred results;
(4) studying on the incremental learning of Bayesian network, and suggestingthat the structure of the Bayesian network should be learned again if the number ofnew data reach10%of the original data;
(5) developing a model for accident duration forecasting based on the theory ofsurvival analysis.
The results can be used to analyze the process of developing and changing oftraffic accident under the influence of the factors and to propose effective trafficsafety management measures. It also provides useful models for prediction of accidentsituation and duration as well as developing rapid response system for traffic accident.The study contributes to the improvement of road traffic safety, reducing the casualtyand property damage caused by traffic accident, and making the social and economicdevelop rapidly.
道路交通事故态势分析的目的是探究道路交通事故与人、车、路、环境等系统之间的相互关系，掌握在各项因素影响下交通事故的发生、发展和生成最终结果的内在规律，以期提出有效的交通安全管理措施，做到在事故发生前预防事故的发生，在事故发生过程中降低事故的严重程度，在事故发生后迅速进行事故快速响应，减少人员伤亡和财产损失。可见，道路交通事故态势研究对于提出并实施行之有效的交通安全管理对策，提高我国的道路交通安全水平，降低交通事故的危害程度，保障社会生产和人民生活，具有非常重要的作用。
本文基于这些背景，结合国家863计划项目《重特大交通事故快速勘测与处置技术》，根据我国的交通事故状况、交通系统特性和社会经济发展特性，建立交通事故态势分析和预测模型，分析各种影响因素与交通事故之间的错综复杂的因果关系，掌握交通事故的发生和发展变化规律，并将研究成果应用于交通安全管理措施的制定和评价以及交通事故快速响应方案的制定，以提高我国的道路交通安全水平，促进道路交通和社会经济的快速发展。
本文完成的主要研究工作有：
论述了交通事故态势分析的基本理论，研究交通事故态势分析的两种主要方法——贝叶斯网络和离散选择模型，并分别应用两种方法建立机动车交通事故的态势分析模型，根据模型检验结果和预测精度进行模型优选；
应用贝叶斯理论，建立非机动车/行人交通事故态势分析模型；
基于已建的机动车和非机动车/行人交通事故态势分析模型，应用团树传播算法，学习各项影响因素对交通事故态势的影响规律；
应用生存分析理论，建立交通事故持续时间预测模型；
根据交通事故态势分析结果，提出降低交通事故严重程度的安全管理措施；
制定交通事故快速响应方案，将已建的交通事故态势分析模型和交通事故持续时间预测模型整合，建立交通事故快速响应系统框架，并研究交通事故态势预测的贝叶斯网络增量学习问题。
本文在以下方面取得了一定创新成果：
构建机动车交通事故和非机动车/行人交通事故态势分析的贝叶斯网络模型，应用贝叶斯网络的结构和参数，对自变量与因变量之间、多维因变量之间以及多个自变量之间的复杂影响关系进行结构化描述和量化表达；
以变量选择的客观性、变量间影响关系的表达、建模精确度等多指标为依据，从定性分析和定量计算两方面对比分析贝叶斯网络模型和离散选择模型在交通事故态势分析方面的优劣；
应用团树传播算法实现贝叶斯网络的推理学习，分析各影响因素的变化对事故严重程度的影响，并将其用于交通事故安全对策的制定和定量评价；
为了解决贝叶斯网络的增量学习问题，提出以新增数据比例10%作为贝叶斯网络结构调整的阈值；
应用生存分析理论构建交通事故持续时间预测的加速失效模型。
本研究成果可用于揭示在各项因素影响下交通事故的发生、发展和生成最终结果的内在规律，提出系统性的交通安全管理措施，预测交通事故态势和持续时间，制定有针对性的交通事故快速响应方案。研究成果对提高我国的道路交通安全水平，降低事故导致的人员伤亡和财产损失，保障社会生产和人民生活，具有重要的理论价值和现实意义。
With the huge progress of China in social, economic, motorization andtransportation recently, the level of road traffic safety is getting lower. Trafficaccidents cause not only casualty and property damage, but also traffic delay and jam,fire, explosion, and leakage of dangerous goods sometimes, which lead to huge lossfor the people. In order to make the social and economic develop rapidly and reducethe damage caused by traffic accident, it is necessary for the government to figure outhow to enhance traffic safety and prevent from traffic accident.
The goal of situation analysis is to examine the relationship among trafficaccident and the potential factors, including people, vehicle, road and environment, aswell as to analyze the process of developing and changing of traffic accident. Basedon situation analysis, effective traffic safety management measures can be made sothat the damage caused by traffic accident can be reduced. Therefore, the researchconcerning about situation analysis is of great importance for enhancing traffic safetyand ensuring the rapid social and economic development.
Founded by National High-tech R&D Program of China (863Program)-QuickInvestigation and Disposal Technologies of Major Accidents, a forecasting model wasdeveloped for traffic accident situation, and the relationship between factors andtraffic accident was analyzed according to the conditions and features of the trafficsystem and social-economic development in China. It then proposed some trafficmanagement measures and developed a program of traffic accident rapid response toimprove traffic safety.
Two potential methods were studied for situation analysis, which are Bayesiannetwork and discrete choice model, and employed them respectively to modelsituation of motor vehicle accident. Bayesian network, which was proved to be betterthan discrete choice model in prediction accuracy, was picked out and employed inmodeling the situations of not only motor vehicle accident, but also non-motorvehicle/pedestrian accident. Then the impacts of the factors on accident situation arelearned by using the Clique Tree Propagation based on the Bayesian networks, andsome traffic management strategies were proposed to reduce damage caused by trafficaccident. A forecasting model of accident duration was also developed according to the theory of survival analysis. Based on the models of situation analysis and durationforecasting, a rapid response system was then developed. The incremental learning ofBayesian network was concerned about at the end of the dissertation.
The major contributions of this dissertation are as follows:
(1) developing situation analysis models for motor vehicle accident andnon-motor vehicle/pedestrian accident, respectively, with Bayesian network, andanalyzing the relationship among dependent variables and independent variables byusing the methods of both structural description and quantitative deduction;
(2) comparing Bayesian network and discrete choice model with respect to thefactors of both qualitative analysis and quantitative calculation, including variableschoosing, relationship analysis and goodness-of-fit of the models, etc.;
(3) inferring the impacts of factors on accident situation with the Clique TreePropagation, and proposing and evaluating traffic management measures according tothe inferred results;
(4) studying on the incremental learning of Bayesian network, and suggestingthat the structure of the Bayesian network should be learned again if the number ofnew data reach10%of the original data;
(5) developing a model for accident duration forecasting based on the theory ofsurvival analysis.
The results can be used to analyze the process of developing and changing oftraffic accident under the influence of the factors and to propose effective trafficsafety management measures. It also provides useful models for prediction of accidentsituation and duration as well as developing rapid response system for traffic accident.The study contributes to the improvement of road traffic safety, reducing the casualtyand property damage caused by traffic accident, and making the social and economicdevelop rapidly.
引文
[1]公安部交通管理局.中华人民共和国道路交通事故统计年报（2000-2011）[R].公安部交通管理局，2001-2011.
[2]吴憩棠.构建中国更安全的道路交通——道路安全国际论坛纪实[J].汽车与配件，2007（31）：16-19.
[3]Gander P E. The impact of speed and other variables on pedestrian safety inMaine[J]. Accident Analysis and Prevention,2004,36(4):533-542.
[4]Harruff R C, Avery A, Alter-Pandya A S. Analysis of circumstances and injuries in217pedestrian traffic fatalities[J].Accident Analysis and Prevention,1998,30(1):11-20.
[5]Atkins R M,Turner W H, Duthie R B, Wilde B R. Injuries to pedestrians in roadTraffic Accidents[J]. British Medical Journal,1988,297(6661):1431-1434.
[6]Jehle D, Cottington E. Effect of alcohol consumption on outcome of pedestrianvictims[J]. Annals of Emergency Medicine,1988,17:953-956.
[7]Holubowycz O T. Age, Sex and blood alcohol concentration of killed and injuredpedestrians[J]. Accident Analysis&Prevention,1995,27(3):417-422
[8]Jensen S U. Pedestrian safety in Denmark[J]. Transportation Research Record,1999,1674:61-69.
[9]Al-Ghamdi A S. Pedestrian-vehicle crashes and analytical techniques for stratifiedcontingency tables[J]. Accident Analysis and Prevention,2002,34:205-214.
[10]Stone M, Broughton J. Getting off your bike: cycling incidents in Great Britain in1990–1999[J]. Accident Analysis and Prevention,2003,35(4):549-556.
[11]Lefler D, Gabler H. The fatality and injury risk of light truck impacts withpedestrians in the United States[J]. Accident Analysis and Prevention,2003,36(2):295-304.
[12]Sze N N, Wong S C. Diagnostic analysis of the logistic model for pedestrianinjury severity in traffic crashes[J]. Accident Analysis&Prevention,2007,39(6):1267-1278.
[13]Kim J K, Kim S, Ulfarsson G F, Porrello L A. Bicyclist injury severities inbicycle–motor vehicle incidents[J]. Accident Analysis and Prevention,2007,39(2):238-251.
[14]Lee C, Abdel-Aty M. Comprehensive analysis of vehicle–pedestrian crashes atintersections in Florida[J]. Accident Analysis and Prevention,2005,37(4):775-786.
[15]Shankar V, Mannering F, Barfield W. Statistical Analysis of Incident Severity onRural Freeways[J]. Accident Analysis and Prevention,1996,28(3):391-401.
[16]Chong Miao, Abraham Ajith. Traffic Accident Analysis using machine learningparadigms[J]. Computational Intelligence in Data Mining,2005,29(5):89-98.
[17]Chong Miao, Abraham Ajith. Traffic Accident Data Mining Using MachineLearning Paradigms[C]. Fourth International Conference on Intelligent SystemsDesign and Applications (ISDA'04), Hungary, ISBN9637154302,2004.
[18]Zhang H, Barbro B, Zhou W. The Design and Implementation of a TrafficAccident Analysis System[C]. Proceedings of the14th International conference onIndustrial and engineering applications of artificial intelligence and expert systems:engineering of intelligent systems,2001:476-481.
[19]Abdelwahab H T, AbdelAty M A. Development of Artificial Neural NetworkModels to Predict Driver Injury Severity in Traffic Accidents at SignalizedIntersections[J]. Transportation Research Record,2001,1746:6-13.
[20]Pitt R, Guyer B, Hsieh C, Malek M. The severity of pedestrian injuries in children:an analysis of the Pedestrian Injury Causation Study[J]. Accident Analysis andPrevention,1990,22(6):549-559.
[21]]Miles-Doan R. Alcohol use among pedestrians and the odds of surviving aninjury: Evidence from Florida Law Enforcement Data[J]. Accident Analysis andPrevention,1996,28(1):23-31.
[22]Klop J R, Khattak A J. Factors influencing bicycle crash severity on two-lane,undivided roadways in North Carolina[J]. Transportation Research Record,1999,1674:78-85.
[23]Ballesteros M F, Dischinger P C, Langenberg P. Pedestrian injuries and vehicletype in Maryland,1995–1999[J]. Accident Analysis and Prevention,2003,36(1):73-81.
[24]Zajac S S, Ivan J N. Factors influencing injury severity of motor vehicle-crossingpedestrian crashes in rural Connecticut[J]. Accident Analysis and Prevention,2003,35(3):369-379.
[25]Roudsari B S, Mock C N, Kaufman R, Grossman D, Henary B Y, Crandall J.Pedestrian crashes: higher injury severity and mortality rate for light truck vehiclescompared with passenger vehicles[J]. Injury Prevention,2004,10:154-158.
[26]李文权，王炜.交通事故的时间分布规律[J].中国安全科学学报，2005,15（4）：56-61.
[27]韩文彬.道路线形与交通事故关系研究[D].长春：吉林大学硕士学位论文，2007.
[28]沙爱敏.高速公路交通事故分析及预防对策研究[D].南京：东南大学硕士学位论文，2006.
[29]魏庆曜，陈斌，金炜东，付锐，袁伟.基于分层关联解析的事故道路因素分析[J].长沙交通学院学报，2005，21（1）：82-86.
[30]魏静.道路交通事故中非主观因素分析及建模[D].天津：天津大学硕士学位论文，2004.
[31]赵杨东.高速公路追尾事故成因分析及预防对策研究[D].哈尔滨：哈尔滨工业大学硕士学位论文，2007.
[32]折欣.道路交通事故致因分析方法研究[D].上海：同济大学硕士学位论文，2008.
[33]李相勇.道路交通事故预测方法研究[D].成都：西南交通大学硕士学位论文，2004.
[34]于泓漪.道路交通事故原因的聚类分析[D].长春：吉林大学硕士学位论文，2005.
[35]缪明月.非机动车驾驶人交通伤害及预防对策研究[J].北京人民警察学院学报，2007，4：46-48.
[36]王宏雁，戎喆慈，朱西产.行人、非机动车道路交通事故特征及伤害研究[J].道路交通与安全，2008，8（3）：28-33.
[37]李松庆.汽车-自行车碰撞事故分析及骑车人头部损伤防护技术研究[D].长沙：湖南大学硕士学位论文，2008.
[38]马壮林，邵春福，李霞.高速公路隧道交通事故严重程度的影响因素分析[J].北京交通大学学报，2009，33（6）：52-55.
[39]马壮林，邵春福，李霞.基于Logistic模型的公路隧道交通事故严重程度的影响因素[J].吉林大学学报（工学版），2010，40（2）：423-426.
[40]李世民，孙明玲，关宏志.基于累积Logistic模型的交通事故严重程度预测模型[J].交通标准化，2009，190（2）：168-171.
[41]姬杨蓓蓓.交通事件持续时间预测方法研究[D].上海：同济大学博士学位论文，2008.
[42]秦小虎.城市交通紧急事件处理与安全系统模型及应用研究[D].重庆：重庆大学博士学位论文，2005.
[43]王发智.基于贝叶斯网络的交通突发事件态势评估技术[D].大连：大连理工大学硕士学位论文，2006.
[44]林震，杨浩.基于车速的交通事故贝叶斯预测[J].中国安全科学学报，2003，13（2）：34-37.
[45]周涛，翟长旭，高志刚.基于贝叶斯网络的高速公路预警系统研究[J].公路工程，2007，32（4）：163-166.
[46]宗芳，张慧永，隽志才.基于贝叶斯网络的停车行为分析[J].系统工程理论与实践，2010，30（5）：948-954.
[47]Kevin Patrick Murphy. Dynamic Bayesian Networks: Representation, Inferenceand learning[D]. Berkeley: University of California,2002.
[48]David Heckerman. A Tutorial on Learning with Bayesian Networks. MicrosoftTechnical Report[R]. No.MSR-TR-95-06. Microsoft Research,1995.
[49]Richard E. Neapolitan. Learning Bayesian Networks[M]. Prentice Hall Series inArtificial Intelligence, Pearson Education, Inc., Pearson Prentice Hall, UpperSaddle River, NJ07458. Chicago: Northeastern Illinois University,2004.
[50]程泽凯.基于TAN结构的启发式贝叶斯网络结构学习算法计算机技术与发展[J].2007，17(8)：61-63.
[51]王利民.贝叶斯学习理论中若干问题的研究[D].长春：吉林大学博士学位论文，2005.
[52]Cooper G, Herskovits E. A Bayesian Method for the Induction of ProbabilisticNetworks from Data[J].Machine Learning,1992,9:309.
[53](美)拉塞尔，(美)诺文著，姜哲等译.人工智能—一种现代方法(第二版)[M].北京：人民邮电出版社，2004.
[54]宗芳，张慧永，贾洪飞.基于贝叶斯网络的停车收费政策评价[J].华南理工大学学报（自然科学版），2010，38（7）：78-83.
[55]张连文，郭海鹏.贝叶斯网引论[M].北京：科学出版社，2006.
[56]李志瑶，宗芳，张屹山.贝叶斯网络推理分析的团树传播算法——以停车行为分析为例[J].长春大学学报，2012，22（5）：505-509.
[57]李志瑶.基于活动的出行需求预测模型研究[D].长春：吉林大学博士学位论文，2006.
[58]杨兆升.交通运输系统规划：有关理论与方法[M].北京：人民交通出版社，1998．
[59]刘灿齐.现代交通规划学[M].北京：人民交通出版社，2001.
[60]宗芳.基于非集计模型的交通需求管理策略评价研究[D].长春：吉林大学博士学位论文，2008.
[61]李强，龙文进.农民工留城与返乡意愿的影响因素分析[J].中国农村经济，2009，2：46-54.
[62]尹希果，王鹏，李刚.城镇水价收费制度研究[J].城市问题，2008，12：73-78.
[63]GA/T16.1-2010道路交通事故信息代码[S]//公安部计算机与信息处理标准化技术委员会.北京：中国标准出版社，2010.
[64]王双成.贝叶斯网络学习、推理和应用[M].上海：立信会计出版社，2010.
[65]Wei C, Lee Y. Sequential forecast of incident duration using Artificial NeuralNetwork models[J]. Accident Analysis& Prevention,2007,39(5):944-954.
[66]Boyles S, Fajardo D, Waller S T. A naive bayesian classifier for incident durationprediction[C]. Proceedings of the TRB86th Annual Meeting Compendium ofPapers, Washington, DC: Springer,2007.
[67]Garib A, Radwan A, Al-Deek H,. Estimating magnitude and duration of incidentdelays[J]. Journal of Transportation Engineering,1997,123(6):459-466.
[68]Ghosh I. Examination of the factors influencing the clearance time of freewayincidents[J]. Journal of Transportation Systems Enginee ring and InformationTechnology,2012,12(3):75-89.
[69]Giuliano G. Incident characteristics, frequency, and duration on a high volumeurban freeway[J]. Transportation Research Part A: General,1989,23(5):387-396.
[70]王文群，陈海波，玛格利特·贝尔.交通事故持续时间分析综述(英文)[J].交通运输系统工程与信息，2005，5(3)：127-14.
[71]赵小强，李瑞敏.城市快速路交通事故持续时间预测[J].公路工程，2010，35(1)：42-44.
[72]Cameron C, Trivedi P K. Microeconometrics: Methods and APPlieations[M].New York: Cambridge University Press,2005:573-600.
[73]Abdulla Alkaabi, Roger Bird, Dilum Dissanayake. Application of Hazard BasedDuration Model to Investigate-Traffic Accident Duration[C].4th IRTADConference, Korea: Seoul,2009,9.
[74]Chung Y. Development of an accident duration prediction model on the koreanfreeway systems[J]. Accident Analysis&Prevention,2010,42(1):282-289.
[75]Chung Y, Yoon B-J. Analytical method to estimate accident duration usingarchived speed profile and its statistical analysis[J]. KSCE Journal of CivilEngineering,2012,16(6):1064-1070.
[76]Golo T F, Recker W W, Leonard J D. An analysis of the severity and incidentduration of truck-involved freeway accidents[J]. Accident Analysis&Prevention,1987,19(5):375-395.
[77]董英荣.通勤者通勤活动安排预测研究[D].长春：吉林大学硕士学位论文，2011.
[78]隽志才，鲜于建川.基于持续时间模型的日出行时间分析[J].交通运输系统工程与信息，2010，10（4）：62-67.
[79]吴文静，隽志才.通勤者下班后非工作活动时间选择决策行为[J].中国公路学报，2010，23（6）：92-95.
[80]Hensher D A, Mannering F L. Hazard-based duration models and their applicationto transport analysis[J]. Transport Reviews,1994,14(1):63-82.
[81]Bhat C R. A generalized multiple durations proportional hazard model with anapplication to activity behavior during the evening work-to-home commute[J].Transportation Research, Part B (Methodological),1996,30B(6):465-480.
[82]Bhat C R, Pinjar A R. Duration modeling. In: Hensher D A, Button K J (Eds.),Handbook of Transport Modelling[M].2nd ed. Amsterdam: Elsevier, Oxford,2008
[83]吴文静.基于时空约束的出行目的地选择行为建模与仿真[D].长春：吉林大学博士学位论文，2011.
[84]Hamed M M, Mannering F L. Modeling travelers’postwork activity involvement:toward a new methodology [J]. Transportation Science,1993,27:381-394.
[85]Abu-Eisheh S A，Mannering F L. Discrete/Continuous analysis of commuters’route and departure time choices[J]. Transportation Research Record,1989,1138:27-34.
[86]Mannering H. Occurrence, frequency, and duration of commuters’ Work-to-homedeparture delay[J]. Transportation Research B,1990,24(2):99-109.
[87]贾红英，王洁贞，赵敬杰.几种生存分析参数模型拟合方法及其应用[J].中国卫生统计，2004，21（4）：201-203.
[88]Colin D. Lewislndustrial and Business Foreeasting Methods[M]. London:Butterworth Scientific,1982.
[89]Chung Y. Development of an incident duration prediction model on the KoreanFreeway Systems[J]. Accident Analysis&Prevention,2009,42(l):282-289.
[90]GB5768-2009道路交通标志和标线[S]//中国国家标准化管理委员会.北京：中国标准出版社，2009.
[91]陈君.道路交通安全管理规划体系及软件研究[D].西安：长安大学硕士学位论文，2005.
[92]周妍.陕西省突发事件应急管理的运行机理及绩效评价研究[D].西安：西安理工大学硕士学位论文，2008.
[93]宗芳.道路交通管理[M].北京：机械工业出版社，2012.
[94]覃容，彭冬芝.事故致因理论探讨[J].华北科技学院学报，2005，2（3）：1-10.
[95]苑春苗，陈宝智，李畅.基于BP神经网络的事故致因分析方法[J].工业安全与环保，2005，31（10）：54-56.
[96]Li Q, Ooeda Y, Matsuaga C, Sumi T. Identifying the determinants of commutingtravel mode choice from daily commutation behavior[J]. Memoirs of the Facultyof Engineering, Kyushu University,2002,62(4):191-206.
[97]Sukumar K, Kara M K. Credit-Based Congestion Pricing: Travel, Land Value andWelfare Impacts[C]. the83rd Annual Meeting of the Transportation ResearchBoard,2004.
[98]Guan H, Sun X, Liu X, Liu L. Modeling Parking Behavior for Better Control andPricing: A Case Study from One of the Busiest Retail Shopping Areas in Beijing,China[C]. the84th Annual Meeting of the Transportation Research Board,2005.
[99]Juan Z, Zong F, Li Z. Study on activity-based time choice model[C]. Chang Chun:Urban Transport and the Environment in the21st Century,2005.
[100]宗芳，隽志才.基于活动的方式选择模型与交通需求管理策略分析[J].吉林大学学报（工学版），2007，37（1）：49-53.
[101]Bowman J L. Activity Based Travel Demand Model System with Daily ActivitySchedules[J]. Master of Science in Transportation at the Massachusettes Instituteof Technology,1995,13(5):56-69.
[102]Jones B, Janssen L, Mannering F. Analysis of the frequency and duration offreeway accidents in seattle[J]. Accident Analysis&Prevention,1991,23(4):239-255.
[103]Khattak A J, Schofer J L, Wang M-H. A simple time sequential procedure forpredicting freeway incident duration[J]. IVHS Journal,1995,2(2):113-138.
[104]Kim H, Choi H. A comparative analysis of incident service time on urbanfreeways[J]. Journal of International Association of Traffic and Safety Sciences,2001,25(1):62-72.
[105]Kim W, Chang G-L. Development of a hybrid prediction model for freewayincident duration: a case study in maryland[J]. International Journal of IntelligentTransportation Systems Research,2012,10(1):22-33.
[106]Kleinbaum D G, Klein M. Survival Analysis: A Self-learning Text[C]. New York:Springer, NY,2012.
[107]Konduri S, Labi S, Sinha K C. Incident occurrence models for freeway incidentmanagement[J]. Transportation Research Record: Journal of the TransportationResearch Board,2003,1856:125-135.
[2]吴憩棠.构建中国更安全的道路交通——道路安全国际论坛纪实[J].汽车与配件，2007（31）：16-19.
[3]Gander P E. The impact of speed and other variables on pedestrian safety inMaine[J]. Accident Analysis and Prevention,2004,36(4):533-542.
[4]Harruff R C, Avery A, Alter-Pandya A S. Analysis of circumstances and injuries in217pedestrian traffic fatalities[J].Accident Analysis and Prevention,1998,30(1):11-20.
[5]Atkins R M,Turner W H, Duthie R B, Wilde B R. Injuries to pedestrians in roadTraffic Accidents[J]. British Medical Journal,1988,297(6661):1431-1434.
[6]Jehle D, Cottington E. Effect of alcohol consumption on outcome of pedestrianvictims[J]. Annals of Emergency Medicine,1988,17:953-956.
[7]Holubowycz O T. Age, Sex and blood alcohol concentration of killed and injuredpedestrians[J]. Accident Analysis&Prevention,1995,27(3):417-422
[8]Jensen S U. Pedestrian safety in Denmark[J]. Transportation Research Record,1999,1674:61-69.
[9]Al-Ghamdi A S. Pedestrian-vehicle crashes and analytical techniques for stratifiedcontingency tables[J]. Accident Analysis and Prevention,2002,34:205-214.
[10]Stone M, Broughton J. Getting off your bike: cycling incidents in Great Britain in1990–1999[J]. Accident Analysis and Prevention,2003,35(4):549-556.
[11]Lefler D, Gabler H. The fatality and injury risk of light truck impacts withpedestrians in the United States[J]. Accident Analysis and Prevention,2003,36(2):295-304.
[12]Sze N N, Wong S C. Diagnostic analysis of the logistic model for pedestrianinjury severity in traffic crashes[J]. Accident Analysis&Prevention,2007,39(6):1267-1278.
[13]Kim J K, Kim S, Ulfarsson G F, Porrello L A. Bicyclist injury severities inbicycle–motor vehicle incidents[J]. Accident Analysis and Prevention,2007,39(2):238-251.
[14]Lee C, Abdel-Aty M. Comprehensive analysis of vehicle–pedestrian crashes atintersections in Florida[J]. Accident Analysis and Prevention,2005,37(4):775-786.
[15]Shankar V, Mannering F, Barfield W. Statistical Analysis of Incident Severity onRural Freeways[J]. Accident Analysis and Prevention,1996,28(3):391-401.
[16]Chong Miao, Abraham Ajith. Traffic Accident Analysis using machine learningparadigms[J]. Computational Intelligence in Data Mining,2005,29(5):89-98.
[17]Chong Miao, Abraham Ajith. Traffic Accident Data Mining Using MachineLearning Paradigms[C]. Fourth International Conference on Intelligent SystemsDesign and Applications (ISDA'04), Hungary, ISBN9637154302,2004.
[18]Zhang H, Barbro B, Zhou W. The Design and Implementation of a TrafficAccident Analysis System[C]. Proceedings of the14th International conference onIndustrial and engineering applications of artificial intelligence and expert systems:engineering of intelligent systems,2001:476-481.
[19]Abdelwahab H T, AbdelAty M A. Development of Artificial Neural NetworkModels to Predict Driver Injury Severity in Traffic Accidents at SignalizedIntersections[J]. Transportation Research Record,2001,1746:6-13.
[20]Pitt R, Guyer B, Hsieh C, Malek M. The severity of pedestrian injuries in children:an analysis of the Pedestrian Injury Causation Study[J]. Accident Analysis andPrevention,1990,22(6):549-559.
[21]]Miles-Doan R. Alcohol use among pedestrians and the odds of surviving aninjury: Evidence from Florida Law Enforcement Data[J]. Accident Analysis andPrevention,1996,28(1):23-31.
[22]Klop J R, Khattak A J. Factors influencing bicycle crash severity on two-lane,undivided roadways in North Carolina[J]. Transportation Research Record,1999,1674:78-85.
[23]Ballesteros M F, Dischinger P C, Langenberg P. Pedestrian injuries and vehicletype in Maryland,1995–1999[J]. Accident Analysis and Prevention,2003,36(1):73-81.
[24]Zajac S S, Ivan J N. Factors influencing injury severity of motor vehicle-crossingpedestrian crashes in rural Connecticut[J]. Accident Analysis and Prevention,2003,35(3):369-379.
[25]Roudsari B S, Mock C N, Kaufman R, Grossman D, Henary B Y, Crandall J.Pedestrian crashes: higher injury severity and mortality rate for light truck vehiclescompared with passenger vehicles[J]. Injury Prevention,2004,10:154-158.
[26]李文权，王炜.交通事故的时间分布规律[J].中国安全科学学报，2005,15（4）：56-61.
[27]韩文彬.道路线形与交通事故关系研究[D].长春：吉林大学硕士学位论文，2007.
[28]沙爱敏.高速公路交通事故分析及预防对策研究[D].南京：东南大学硕士学位论文，2006.
[29]魏庆曜，陈斌，金炜东，付锐，袁伟.基于分层关联解析的事故道路因素分析[J].长沙交通学院学报，2005，21（1）：82-86.
[30]魏静.道路交通事故中非主观因素分析及建模[D].天津：天津大学硕士学位论文，2004.
[31]赵杨东.高速公路追尾事故成因分析及预防对策研究[D].哈尔滨：哈尔滨工业大学硕士学位论文，2007.
[32]折欣.道路交通事故致因分析方法研究[D].上海：同济大学硕士学位论文，2008.
[33]李相勇.道路交通事故预测方法研究[D].成都：西南交通大学硕士学位论文，2004.
[34]于泓漪.道路交通事故原因的聚类分析[D].长春：吉林大学硕士学位论文，2005.
[35]缪明月.非机动车驾驶人交通伤害及预防对策研究[J].北京人民警察学院学报，2007，4：46-48.
[36]王宏雁，戎喆慈，朱西产.行人、非机动车道路交通事故特征及伤害研究[J].道路交通与安全，2008，8（3）：28-33.
[37]李松庆.汽车-自行车碰撞事故分析及骑车人头部损伤防护技术研究[D].长沙：湖南大学硕士学位论文，2008.
[38]马壮林，邵春福，李霞.高速公路隧道交通事故严重程度的影响因素分析[J].北京交通大学学报，2009，33（6）：52-55.
[39]马壮林，邵春福，李霞.基于Logistic模型的公路隧道交通事故严重程度的影响因素[J].吉林大学学报（工学版），2010，40（2）：423-426.
[40]李世民，孙明玲，关宏志.基于累积Logistic模型的交通事故严重程度预测模型[J].交通标准化，2009，190（2）：168-171.
[41]姬杨蓓蓓.交通事件持续时间预测方法研究[D].上海：同济大学博士学位论文，2008.
[42]秦小虎.城市交通紧急事件处理与安全系统模型及应用研究[D].重庆：重庆大学博士学位论文，2005.
[43]王发智.基于贝叶斯网络的交通突发事件态势评估技术[D].大连：大连理工大学硕士学位论文，2006.
[44]林震，杨浩.基于车速的交通事故贝叶斯预测[J].中国安全科学学报，2003，13（2）：34-37.
[45]周涛，翟长旭，高志刚.基于贝叶斯网络的高速公路预警系统研究[J].公路工程，2007，32（4）：163-166.
[46]宗芳，张慧永，隽志才.基于贝叶斯网络的停车行为分析[J].系统工程理论与实践，2010，30（5）：948-954.
[47]Kevin Patrick Murphy. Dynamic Bayesian Networks: Representation, Inferenceand learning[D]. Berkeley: University of California,2002.
[48]David Heckerman. A Tutorial on Learning with Bayesian Networks. MicrosoftTechnical Report[R]. No.MSR-TR-95-06. Microsoft Research,1995.
[49]Richard E. Neapolitan. Learning Bayesian Networks[M]. Prentice Hall Series inArtificial Intelligence, Pearson Education, Inc., Pearson Prentice Hall, UpperSaddle River, NJ07458. Chicago: Northeastern Illinois University,2004.
[50]程泽凯.基于TAN结构的启发式贝叶斯网络结构学习算法计算机技术与发展[J].2007，17(8)：61-63.
[51]王利民.贝叶斯学习理论中若干问题的研究[D].长春：吉林大学博士学位论文，2005.
[52]Cooper G, Herskovits E. A Bayesian Method for the Induction of ProbabilisticNetworks from Data[J].Machine Learning,1992,9:309.
[53](美)拉塞尔，(美)诺文著，姜哲等译.人工智能—一种现代方法(第二版)[M].北京：人民邮电出版社，2004.
[54]宗芳，张慧永，贾洪飞.基于贝叶斯网络的停车收费政策评价[J].华南理工大学学报（自然科学版），2010，38（7）：78-83.
[55]张连文，郭海鹏.贝叶斯网引论[M].北京：科学出版社，2006.
[56]李志瑶，宗芳，张屹山.贝叶斯网络推理分析的团树传播算法——以停车行为分析为例[J].长春大学学报，2012，22（5）：505-509.
[57]李志瑶.基于活动的出行需求预测模型研究[D].长春：吉林大学博士学位论文，2006.
[58]杨兆升.交通运输系统规划：有关理论与方法[M].北京：人民交通出版社，1998．
[59]刘灿齐.现代交通规划学[M].北京：人民交通出版社，2001.
[60]宗芳.基于非集计模型的交通需求管理策略评价研究[D].长春：吉林大学博士学位论文，2008.
[61]李强，龙文进.农民工留城与返乡意愿的影响因素分析[J].中国农村经济，2009，2：46-54.
[62]尹希果，王鹏，李刚.城镇水价收费制度研究[J].城市问题，2008，12：73-78.
[63]GA/T16.1-2010道路交通事故信息代码[S]//公安部计算机与信息处理标准化技术委员会.北京：中国标准出版社，2010.
[64]王双成.贝叶斯网络学习、推理和应用[M].上海：立信会计出版社，2010.
[65]Wei C, Lee Y. Sequential forecast of incident duration using Artificial NeuralNetwork models[J]. Accident Analysis& Prevention,2007,39(5):944-954.
[66]Boyles S, Fajardo D, Waller S T. A naive bayesian classifier for incident durationprediction[C]. Proceedings of the TRB86th Annual Meeting Compendium ofPapers, Washington, DC: Springer,2007.
[67]Garib A, Radwan A, Al-Deek H,. Estimating magnitude and duration of incidentdelays[J]. Journal of Transportation Engineering,1997,123(6):459-466.
[68]Ghosh I. Examination of the factors influencing the clearance time of freewayincidents[J]. Journal of Transportation Systems Enginee ring and InformationTechnology,2012,12(3):75-89.
[69]Giuliano G. Incident characteristics, frequency, and duration on a high volumeurban freeway[J]. Transportation Research Part A: General,1989,23(5):387-396.
[70]王文群，陈海波，玛格利特·贝尔.交通事故持续时间分析综述(英文)[J].交通运输系统工程与信息，2005，5(3)：127-14.
[71]赵小强，李瑞敏.城市快速路交通事故持续时间预测[J].公路工程，2010，35(1)：42-44.
[72]Cameron C, Trivedi P K. Microeconometrics: Methods and APPlieations[M].New York: Cambridge University Press,2005:573-600.
[73]Abdulla Alkaabi, Roger Bird, Dilum Dissanayake. Application of Hazard BasedDuration Model to Investigate-Traffic Accident Duration[C].4th IRTADConference, Korea: Seoul,2009,9.
[74]Chung Y. Development of an accident duration prediction model on the koreanfreeway systems[J]. Accident Analysis&Prevention,2010,42(1):282-289.
[75]Chung Y, Yoon B-J. Analytical method to estimate accident duration usingarchived speed profile and its statistical analysis[J]. KSCE Journal of CivilEngineering,2012,16(6):1064-1070.
[76]Golo T F, Recker W W, Leonard J D. An analysis of the severity and incidentduration of truck-involved freeway accidents[J]. Accident Analysis&Prevention,1987,19(5):375-395.
[77]董英荣.通勤者通勤活动安排预测研究[D].长春：吉林大学硕士学位论文，2011.
[78]隽志才，鲜于建川.基于持续时间模型的日出行时间分析[J].交通运输系统工程与信息，2010，10（4）：62-67.
[79]吴文静，隽志才.通勤者下班后非工作活动时间选择决策行为[J].中国公路学报，2010，23（6）：92-95.
[80]Hensher D A, Mannering F L. Hazard-based duration models and their applicationto transport analysis[J]. Transport Reviews,1994,14(1):63-82.
[81]Bhat C R. A generalized multiple durations proportional hazard model with anapplication to activity behavior during the evening work-to-home commute[J].Transportation Research, Part B (Methodological),1996,30B(6):465-480.
[82]Bhat C R, Pinjar A R. Duration modeling. In: Hensher D A, Button K J (Eds.),Handbook of Transport Modelling[M].2nd ed. Amsterdam: Elsevier, Oxford,2008
[83]吴文静.基于时空约束的出行目的地选择行为建模与仿真[D].长春：吉林大学博士学位论文，2011.
[84]Hamed M M, Mannering F L. Modeling travelers’postwork activity involvement:toward a new methodology [J]. Transportation Science,1993,27:381-394.
[85]Abu-Eisheh S A，Mannering F L. Discrete/Continuous analysis of commuters’route and departure time choices[J]. Transportation Research Record,1989,1138:27-34.
[86]Mannering H. Occurrence, frequency, and duration of commuters’ Work-to-homedeparture delay[J]. Transportation Research B,1990,24(2):99-109.
[87]贾红英，王洁贞，赵敬杰.几种生存分析参数模型拟合方法及其应用[J].中国卫生统计，2004，21（4）：201-203.
[88]Colin D. Lewislndustrial and Business Foreeasting Methods[M]. London:Butterworth Scientific,1982.
[89]Chung Y. Development of an incident duration prediction model on the KoreanFreeway Systems[J]. Accident Analysis&Prevention,2009,42(l):282-289.
[90]GB5768-2009道路交通标志和标线[S]//中国国家标准化管理委员会.北京：中国标准出版社，2009.
[91]陈君.道路交通安全管理规划体系及软件研究[D].西安：长安大学硕士学位论文，2005.
[92]周妍.陕西省突发事件应急管理的运行机理及绩效评价研究[D].西安：西安理工大学硕士学位论文，2008.
[93]宗芳.道路交通管理[M].北京：机械工业出版社，2012.
[94]覃容，彭冬芝.事故致因理论探讨[J].华北科技学院学报，2005，2（3）：1-10.
[95]苑春苗，陈宝智，李畅.基于BP神经网络的事故致因分析方法[J].工业安全与环保，2005，31（10）：54-56.
[96]Li Q, Ooeda Y, Matsuaga C, Sumi T. Identifying the determinants of commutingtravel mode choice from daily commutation behavior[J]. Memoirs of the Facultyof Engineering, Kyushu University,2002,62(4):191-206.
[97]Sukumar K, Kara M K. Credit-Based Congestion Pricing: Travel, Land Value andWelfare Impacts[C]. the83rd Annual Meeting of the Transportation ResearchBoard,2004.
[98]Guan H, Sun X, Liu X, Liu L. Modeling Parking Behavior for Better Control andPricing: A Case Study from One of the Busiest Retail Shopping Areas in Beijing,China[C]. the84th Annual Meeting of the Transportation Research Board,2005.
[99]Juan Z, Zong F, Li Z. Study on activity-based time choice model[C]. Chang Chun:Urban Transport and the Environment in the21st Century,2005.
[100]宗芳，隽志才.基于活动的方式选择模型与交通需求管理策略分析[J].吉林大学学报（工学版），2007，37（1）：49-53.
[101]Bowman J L. Activity Based Travel Demand Model System with Daily ActivitySchedules[J]. Master of Science in Transportation at the Massachusettes Instituteof Technology,1995,13(5):56-69.
[102]Jones B, Janssen L, Mannering F. Analysis of the frequency and duration offreeway accidents in seattle[J]. Accident Analysis&Prevention,1991,23(4):239-255.
[103]Khattak A J, Schofer J L, Wang M-H. A simple time sequential procedure forpredicting freeway incident duration[J]. IVHS Journal,1995,2(2):113-138.
[104]Kim H, Choi H. A comparative analysis of incident service time on urbanfreeways[J]. Journal of International Association of Traffic and Safety Sciences,2001,25(1):62-72.
[105]Kim W, Chang G-L. Development of a hybrid prediction model for freewayincident duration: a case study in maryland[J]. International Journal of IntelligentTransportation Systems Research,2012,10(1):22-33.
[106]Kleinbaum D G, Klein M. Survival Analysis: A Self-learning Text[C]. New York:Springer, NY,2012.
[107]Konduri S, Labi S, Sinha K C. Incident occurrence models for freeway incidentmanagement[J]. Transportation Research Record: Journal of the TransportationResearch Board,2003,1856:125-135.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |