摘要
提出了一种改进的基于密度和网格的高维聚类算法,并对算法有效性进行了验证。该算法通过减少样本点数量的方法达到减少稠密子空间数量。在发现高维稠密子空间时,对样本库进行精简。这些样本点的求得能有效减少求解最小聚类的时间复杂度。
This paper proposes an improved high- dinaensional cluster analysis algorithm based on grid and intensity , then discusses it＇s validity validation. The amount of the density subspace can be deduced by cutting down that of sample data . The sampie library is simplified as the high - dimensional subspaces are found. By working out such sample data the time complexity of figuring out min cluster is effectively reduced.
出处
《舰船电子工程》
2005年第5期55-56,59,共3页
Ship Electronic Engineering
关键词
数据挖掘
聚类
网格
密度
高维数据
子空间
最小聚类
data mining, cluster, grid, density, high - dimensional data, subspace, min cluster
