基于高斯混合聚类模型的公交出行特征分析

Analysis of bus trip characteristics based on Gaussian mixture clustering model

针对公交出行特征的传统数据分析方法人工成本大的问题,提出一种基于高斯混合聚类模型的公交出行特征分析方法。以公交IC卡刷卡数据、公交运行GPS数据及静态站点数据为基础,建立高斯混合聚类模型,对比节假日与通勤日公交出行特征差异。最后以深圳市某路公交为实例,仿真结果表明,节假日与通勤日公交出行在高峰时段分布与持续时间上具有显著差异,验证了高斯混合聚类模型在交通数据分析领域中的有效性,对公交运营与调度优化有一定的借鉴意义。

In allusion to the problem of large labor cost for traditional data analysis methods of bus trip characteristics,an analysis method of bus travel characteristics based on Gaussian mixture clustering model is proposed.On the basis of bus IC card swiping data,bus running GPS data and static station site data,a Gaussian mixture clustering model was established to compare the differences in bus trip characteristics of holidays and commuting days.A simulation experiment was performed by taking a road bus in Shenzhen city as an example.The simulation results show that there are significant differences in the distri-bution and duration of bus trip in holidays and commuting days during peak hours,which verifies the effectiveness of Gaussian mixture clustering model in the field of transportation data analysis,and has certain reference significance for bus operation and scheduling optimization.