居民出行与轨迹行为交互模式挖掘与关联技术

Research on the interactive mode of residents’behavior based on trajectory data mining

出租车是居民出行的重要交通工具,其轨迹数据蕴含着丰富的居民出行信息。原始出租车轨迹数据因缺少语义信息无法直观反映居民出行规律。通过轨迹数据挖掘技术处理之后的出租车轨迹数据能够反映居民活动规律和行为模式,从而为城市规划决策提供参考依据。本文重点研究了基于语义的交互模式度量,通过出租车停留点推断其语义信息;然后根据语义信息构建语义交互矩阵,用以推断和描述行为目的交互模式;最后选取北京市中心为研究区域进行方法验证。结果表明,中心城区内不同类别的停留点聚集分布规律不同,围绕高校和商圈聚集较明显;工作日各类停留点的活跃度持续时间较非工作日长;工作日和非工作日行为目的交互模式差别显著,工作日以职住和工作交互为主,非工作日以休闲和居住交互为主。本文研究可以为城市规划管理、资源调度和应急管理提供一定的决策支持。

As one of the effective means of transportation,taxis carry much valuable residents’traveling information in their trajectory data.However,unprocessed trajectory data is incapable of revealing the information due to the lack of semantic information.Trajectory,data mining technology,provides a method to unveil residents activating regular patterns,which can support urban planning.This paper provides an interactive mode measuring method based on semantic information to mine the implicit relationship between residents’behavioral purpose and trajectory.Firstly,the author discovers the semantic information from the location of the stay points.Then constructs the Interactive matrix of semantic to infer the interaction mode of behavior purpose.Finally,it takes the central area of Beijing as the research area to test the method.The test shows the association of the property of stay points and their distribution,which is in accord with the reality:here are apparent clusters around universities and business districts;the active of stay points is more durable on weekdays than on the weekends.There is a significant difference in the interactive mode between weekdays and weekends:the primary interactive mode is dwelling-working on the weekdays and is dwelling-relaxing on the weekend.This research can provide an essential reference for urban planning,resource allocation,and emergency management.