利用手机信令数据刻画不同人物画像

Portraying of different characters based on mobile phone signaling data

本文基于微软亚洲研究院Geolife项目北京志愿者2007—2012年长时间尺度的手机信令数据,以个体为单元开展了出行类型和人物画像研究,提出了一种基于高簇聚类对用户轨迹类型进行划分,然后结合出行规律,综合考虑职业类型、年龄、爱好属性特征的人物画像刻画的方法。研究主要结论为:①本文划分了包括两点一线固定型、两点一线变化型、双核心型、均匀分布型、发散型5种出行类型;②北京市拥有固定工作的技术人员或白领人群约占44%,学生群体或退休老人有接近1/4的数量。

Based on the long-term mobile phone signaling data of Beijing volunteers from Microsoft Research Asia Geolife project from2007 to 2012,this article conducts research on travel types and portraits based on individuals. This paper proposes a method of character portrait characterization by first classifying user trajectory types based on high-cluster clustering,and then combining travel rules to consider occupation type,age,and hobby attributes. The main conclusions of the study are: ① This paper divides five types of travel including two-point,one-line fixed type,two-point,one-line change type,dual-core type,uniform distribution type,and divergent type. ② Beijing has fixed-work technology people or white-collar workers account for about 44%,and the student group or retired elderly account for nearly a quarter.