摘要
对交通出行数据进行优化,抽取出租车载客过程中乘客上下车的GPS位置坐标。基于聚类与交通小区划分的相似性,采用K-Means聚类法进行交通小区的划分。首先,通过聚类得到交通出行OD矩阵,然后据此划分出交通小区。基于GoogleMapsAPI,搭建了软件平台。通过试验可以看出,这种动态划分方法得到的区域能够与现有的交通小区相吻合。这种高实时的交通小区划分方法将对动态的OD估计有着极大的参考价值。
The paper optimizes traffic travel data by extracting the Taxi GPS position coordinates of the passenger boarding and alight spots. The paper employs K-means cluster analysis to the partition of traffic zones. First it obtains the traffic trip OD matrix through clustering, which is then used as reference for the partitioning of traffic zones. It has also set up a software platform based on Google Maps API. Finally, through a experimental study, the dynamic partition method is proven to be able to yield outcomes conforming to the existing traffic zones.
出处
《物流技术》
2010年第9期86-88,135,共4页
Logistics Technology
基金
国家科技支撑计划项目(2007BAK12B04-15)
国家"863"高科技资助项目(2006AA11Z231)
北京市科技计划重点项目(D07020601400707)
北京市教育委员会学科建设与研究生教育建设项目
关键词
交通小区
动态分析
K均值聚类
边界计算
traffic zone
dynamic analysis
K-means cluster
boundary calculation
