出租车轨迹数据的南京人群出行模式挖掘

Movement pattern mining of Nanjing residents based on taxi trajectory data

针对现有出租车轨迹数据挖掘中时间序列邻近度量方法存在的问题,提出一种基于DBSCAN算法和改进的DTW距离的时间序列聚类算法提取具有相似性出行特征的时空模式,进而研究城市人群出行行为的时空差异。以南京市为例,结合电子地图对出行模式的空间分布特征进行分析,证明了本文所提出的方法的有效性。实验结果表明:在空间分布上,工作日出租车出行模式按照平均出行频次由高到低排序,从城市中心向四周扩散,呈中心环状分布,出行模式区域界限较为明显,同类出行模式分布区域对应相似的功能。提出了一种基于DBSCAN算法和改进的DTW距离的时间序列聚类算法提取具有相似性出行特征的时空模式,有效地分析城市人群出行行为的时空差异。

Aiming at the problems of time series proximity measurement methods in existing taxi trajectory data mining,a time series clustering algorithm based on DBSCAN algorithm and improved DTW distance is proposed to extract spatiotemporal patterns with similar travel characteristics,and then to study temporal and spatial differences in travel behavior of urban crowd.Taking Nanjing city as an example,the analysis of the spatial distribution characteristics of travel patterns combined with electronic maps proves the effectiveness of the method proposed in this article.The results show that human mobility patterns in working days is arranged from high to low according to the average travel frequency,and spreads from the city center to the surrounding area,showing a central circular distribution.The distribution of movement patterns is influenced by the type of land use or urban functions.A time series clustering algorithm based on DBSCAN algorithm and improved DTW distance is proposed to extract spatio-temporal patterns with similar travel characteristics,and effectively analyze the spatiotemporal differences of urban crowd travel behavior.