分布式增量机制下的交通流大数据聚类分析

Traffic Flow Big Data Clustering Analysis Method Based on Distributed Incremental Mechanism

时空聚类分析是对时空大数据进行利用的一种有效手段。本文提出了一种分布式增量大数据聚类分析方法,利用分布增量机制不但可以减少重复计算和迁移拷贝次数,而且可以持续对聚类结果进行修正,能够在保持聚类准确性的条件下提升整体运算效率。而聚类算法本身通过数据聚集趋势预分析、聚类算法和结果评价3个步骤,构建了一体化时空邻域,在时间和空间维度保证了聚类结果的准确性。经过试验证明该方法可以实现时空大数据的快速高效信息挖掘。

Spatio-temporal clustering analysis is an effective way of using spatio-temporal big data. This paper proposes a distributed incremental big data clustering analysis method. The incremental distribution mechanism can not only reduce the repeated calculation and the number of copies, but also can modify the clustering results continuously. And it is able to improve the operational efficiency under the condition of keeping in clustering accuracy. The clustering algorithm includes three steps：data aggregation trend analysis, clustering algorithm and result evaluation. It constructs an integrated spatio-temporal neighborhood, which guarantees the accuracy of clustering results in time and space. The experiments show that this method can realize the fast and efficient information mining of spatio-temporal big-data.