项目名称: 人类移动行为的统计特征分析与预测方法研究
项目编号: No.61304177
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 闫小勇
作者单位: 北京交通大学
项目金额: 23万元
中文摘要: 人类移动行为复杂性研究是当前复杂系统研究领域的一个重要主题。理解和预测人的移动行为对于研究受人类活动影响的各种复杂现象具有重要理论意义,在疾病传播防控、交通规划、商业服务等领域也具有广泛应用价值。本项目在收集大规模人类移动行为数据的基础上,深入挖掘人类移动行为的时空统计特征,分别建立面向个体移动轨迹预测和群体移动模式预测的模型。在统计特征分析方面,主要研究个体移动轨迹的周期性、规则性等特征以及群体移动模式与城市人口分布、交通网络结构等因素之间的相互关系;在个体移动轨迹预测方法研究方面,充分利用移动轨迹中蕴含的时空结构信息,建立低数据需求和高精度的轨迹预测模型;在群体移动模式预测方法研究方面,基于热传导理论建立能够准确预测城市人群移动量的无参数模型。本项目研究结果可深化人类对自身移动行为规律的理解,丰富和发展人类移动行为复杂性研究的理论与方法。
中文关键词: 人类动力学;人类移动模式;标度律;最大熵;出行预测
英文摘要: Research on the complexity of human mobility behaviors is an important issue in complexity sciences. Understanding and predicting human mobility patterns are of importance for understanding some complex systems affected by human activities and have many practical applications in epidemiology of infectious disease, transportation planning, location-based service and so on. In this project, we will explore the spatio-temporal statistical properties of human mobility behaviors and develop practical models for predicting individual mobility trajectories and crowd mobility patterns based on the big data of real-life human mobility records. The research content of the project mainly includes: (i) empirically analysing the periodicity and regularity of individual human mobility trajectories and the relationships between the crowd mobility patterns and their influential factors such as urban population distribution and transportation networks structure; (ii) developing high accurate and low data requirement methods to predict the individual mobility trajectories by using the spatio-temporal structural information embedded in human daily travel activity sequences; (iii) developing a parameter-free and high accurate model to predict crowd mobility patterns in cities based on heat conduction theory. The results of the proj
英文关键词: human dynamics;human mobility patterns;scaling law;maximum entropy;travel prediction