利用优化的DenStream算法进行空间数据流聚类  

Spatial Data Stream Clustering by Optimized DenStream Algorithm

作　　者：樊超 李宏伟[1] 

机构地区：[1]信息工程大学,河南郑州450001

出　　处：《测绘与空间地理信息》2017年第4期73-77,共5页Geomatics & Spatial Information Technology

基　　金：空间数据流的概念漂移问题研究(41571394)资助

摘　　要：为了有效解决DenStream算法在空间数据流聚类应用中存在的密度空间分布不均的问题,本文提出使用相对密度比代替密度作为聚类参数,通过考虑微簇周围密度环境,降低密度分布不均对聚类的影响。同时,使用空间格网索引,方便查找周围的微簇与数据点,进而提高算法效率。最后,通过使用真实数据对优化前后的算法进行对比,验证了优化后的算法在继承DenStream算法优点的基础上,有效地避免了密度空间分布不均的问题。In order to effectively avoid the problem that DenStream algorithm has the problem of spatial distribution of spatial densityin spatial data stream clustering, it is proposed to use relative density ratio instead of density as clustering parameter and reduce thedensity distribution inequality by clustering Impact. At the same time, using the spatial grid index, easy to find around the cluster anddata points, thereby improving the efficiency of the algorithm. Finally, the comparison of the algorithms before and after the optimiza-tion is carried out by using the real data. It is verified that the optimized algorithm inherits the advantages of DenStream algorithm, andavoids the problem of uneven density distribution.

关 键 词：空间数据挖掘 空间数据流 DenStream算法 聚类 密度 

分 类 号：P209[天文地球—测绘科学与技术]