摘要
提出了一种基于空间单元单维运算的快速聚类算法SUSDC。该算法首先将被聚类的数据逐维划分成若干个不相交的空间单元;然后基于空间距离阈值判定相邻的空间单元是否合并,直到全部维处理完毕。实验结果验证了SUSDC算法运算速度快,能够处理不规则形状数据和高维数据,且具有对噪声数据不敏感的特点。
This paper proposed a rapid clustering algorithm based on spatial unit single dimensional calculation (SUSDC). The algorithm firstly divided data to orthogonal spatial units, and then clustered spatial units by spatial distance threshold. Experimental results show that SUSDC algorithm can deal with clusters of complicated shaped and high-dimensional data, and it has high-efficient features and is not sensitive to noise data.
出处
《计算机应用研究》
CSCD
北大核心
2009年第8期2858-2860,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(60736045)
关键词
聚类
空间聚类
单维距离
空间密度
clustering
spatial clustering
single dimensional distance
spatial density