一种基于网格和距离阈值的空间数据流聚类算法  

Clustering Algorithm for Spatial Data Streams based on A Grid and Distance Threshold

作　　者：闫昭博[1] 赵静[2] 朱丽萍[2] 

机构地区：[1]武汉大学计算机学院,武汉430072 [2]武汉大学资源与环境学院,武汉430072

出　　处：《山西大学学报（自然科学版）》2008年第4期521-525,共5页Journal of Shanxi University(Natural Science Edition)

摘　　要：空间数据库中存储了海量的实时数据,常规的算法在数据挖掘过程中无法根据空间数据流的特征进行聚类.针对该问题,我们提出一种基于网格和距离阈值的空间数据流聚类算法,该算法分为离线和在线部分,为实时数据赋予时间相关的距离值,利用网格进行聚类.实验证明此算法对于空间数据流聚类具有理想的结果.Spatial database stores huge volumes of real-time data. The conventional method can not cluster on the basis of the characteristics of spatial data streams in the data mining process. In order to address the problem,a spatial data streams clustering algorithm based on grid and distance threshold is presented. The algorithm is divided into the offline and online part. It assigns a distance value related with the time to the real-time data,using grid to cluster. It＇s proved that the algorithm for the spatial data streams clustering is satisfied by experiments.

关 键 词：网格 距离阈值 空间数据流 聚类 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]