基于智能刷卡数据的地铁乘客分类研究

Research on Classification of Metro Passengers Based on Smart Card Data

为了解决根据自身属性对乘客分类存在主观性较强的问题,提出一种基于出行特征的地铁乘客分类方法。以杭州市地铁智能刷卡数据为例,以用户卡号为索引提取个体出行轨迹,构建包含进站时间、进站线路、进站站点、出站时间、出站线路和出站站点的用户出行链,从中提取乘客出行强度特征、时间特征和空间特征,利用二阶聚类算法建立乘客出行的分层聚类模型,首先基于出行强度特征对乘客进行初始层聚类,然后基于时间特征和空间特征进行第二层聚类,最终将乘客分为8类,并分析不同类别乘客的出行规律及总体特征。

In order to solve the problem of strong subjectivity in passenger classification based on its own attributes,a method of subway passenger classification based on travel characteristics is proposed. Taking the smart card swipe data of Hangzhou Metro as an example,the individual travel trajectory is extracted with the user card number as an index;and a user travel chain including the inbound time,inbound line,inbound station,outbound time,outbound line,and outbound station are constructed. The characteristics of passenger travel intensity,time characteristics and space characteristics are extracted;and the second-order clustering algorithm is used to establish a hierarchical clustering model of passenger travel. Firstly,the passengers are clustered initially based on the travel intensity characteristics,and then based on the time characteristics and spatial characteristics. The second level of clustering finally passengers are divided into 8 categories;and the travel rules and overall characteristics of different categories of passengers are analyzed.