一种对二维空间对象进行聚类的算法

An Algorithm for Clustering 2-D Spatial Objects

地理信息系统存储了大量的二维空间对象,对这些对象进行聚类分析是数据挖掘的一项重要任务。本文提出一种针对二维空间对象的聚类算法,该算法引用层次聚类方法的思想,将子聚类信息用一个聚类特征表示。采用基于密度的方法,发现任意形状的簇,能较好地处理孤立点,并且支持增量式聚类。实验证明该算法是有效的。

Many 2--D objects have been stored in geographic information system. Clustering those objects is an important task of Data Mining. This paper proposes an density -- based algorithm especially for clustering 2--D spatial objects. By taking the advantage of hierarchical clustering algorithm, it depicts the information of a cluster with a cluster -- feature. Based on density, this algorithm can discover clusters with any shape, identify noises and support incremental clustering. Experiments demonstrate that this algorithm is effective.