摘要
随着交通规划由注重设施建设向注重居民出行需求管理转变,非汇总层面的居民时空行为模式研究变得尤为重要,基于GPS、GSM等设备获取的个体行为时空数据也为居民活动特征研究提供了新契机。如何基于个体历史活动时空间数据,把握居民活动规律性,进而实现对居民活动行为时空间预测成为学界研究的热点。该文以北京市为案例,通过GPS对2012年北京居民的跟踪调查数据,基于信息熵算法和费诺不等式算法,探究居民个体活动行为的可预测性。研究表明:居民活动行为的可预测性与活动半径存在显著的相关关系,活动半径在10km以内的居民移动行为可预测性普遍在92%以上;居民一日之内的活动行为可预测性具有明显的时间节律,呈现双波谷的变化特征。
With the development of information technology and the spread of humanism thought, the transportation planning＇s focus has turned from infrastructure instruction to residents travel demand management, and the research on residents＇ activity characteristics of aggregate level becomes especially important. Along with the development and wide application of ICT, it is possible to acquire detailed individual behavior time-space data based on GPS and GSM, and also provides a new opportunity for residents＇ activity characteristics researctL How to individual history activity time-space data, to grasp the residents＇ activity regularity, thus realize time-space prediction of the residents＇ activity becomes the focus of academic research. Using information entropy and leers inequality based on a GPS-facilitated activity-travel survey dataset collected in Beijing in 2012, this research explores the predictability of residents＇ individual activity behavior. It is indicated that residentsp behavior can be highly predicted, the predictability of residents within 10 km radius is generally above 92%. Further analysis finds that the predictability significantly relevant to radius of activity, and the predictability within one day has obvious time rhythm, showing the double troughs features. The paper is an effective attempt in the field of resident travel demand management application, provides a scientific basis for precise mobile behavior prediction, and has essential significance on the behavior analysis of time and space, human movement patterns and urban planning and management.
出处
《地理与地理信息科学》
CSCD
北大核心
2015年第6期47-51,共5页
Geography and Geo-Information Science
基金
"十二五"国家科技计划项目课题(2012BAJ05B04)
