Abstract: In order to explore the influence range of the abnormal state of multi-mode transport system in urban agglomeration, the abnormal environments in multimodal transportation system were classified from two aspects: abnormal changes in passenger flow on the demand side and a decrease in transportation capacity on the supply side.A process and analysis method were proposed for identifying the influence range of abnormal states based on multi-source data such as mobile signaling data, intention to travel surveys and passenger ticket information.Taking the Beijing-Tianjin-Hebei urban agglomeration as an example, the dynamic safety threshold of passenger flow between strong link hubs was determined, and an abnormal detection method of channel passenger flow was proposed based on Bayesian prediction.The effectiveness of hub association rules was distinguished by lifting degree, cosine similarity and their normalized values, and a method to determine the influence range of abnormal states based on strong association rules was proposed.The research results show that the channel passenger flow abnormal detection method based on Bayesian prediction has real-time performance and improves the accuracy of passenger flow safety threshold.The influence scope of the hubs is determined by the actual mobile phone data, the intention survey data under the assumed abnormal conditions as well as the association rules, which can provide reference for using data mining technology to analyze the impact and countermeasures of sudden abnormal states.

Key words: traffic and transportation engineering; comprehensive transportation; urban agglomeration; abnormal state; influence range; association rules; multi-source data

摘要： 为探究城市群多模式交通系统异常状态的影响范围，从需求端非常态客流变化和供给端运输能力下降两方面对多模式交通系统异常环境进行分类，提出了基于手机信令数据、意向出行调查及客票信息等多源数据的异常状态影响范围识别流程和分析方法；以京津冀城市群为例，确定强链接枢纽间通道客流动态安全阈值，提出了基于贝叶斯预测的通道客流异常检测方法；利用提升度、余弦相似度及其标准化值来区分枢纽关联规则的有效性，提出了基于强关联规则确定异常状态影响范围的方法。研究表明：基于贝叶斯预测的通道客流异常检测方法具有实时性，提高了客流安全阈值精度；基于实际手机数据和假定异常情况下的意向调查数据并结合关联规则确定枢纽影响范围，可为利用数据挖掘技术分析突发异常状态的影响和对策提供借鉴。

关键词: 交通运输工程；综合交通运输；城市群；异常状态；影响范围；关联规则；多源数据