摘要
聚类是数据挖掘中的关键问题,吸取了BIRCH算法中构造簇特征树来产生初始聚类中心的方法,提出了一种基于随机抽取的有限深度层次聚类算法(RSLDCH算法),采用随机抽取样本、限制特征树深度、构建叶子节点链表技术从而提高了算法的时间效率和聚类效果.实验表明,RSLDCH较BIRCH在运行速度和聚类效果上有一定的提高.
Clustering is a key problem in data mining.A hierarchical clustering algorithm is proposed based on random sampling and limited depth,which absorbs the method BIRCH algorithm used constructing a clustering feature tree to induce a group of initial clustering centers.Simultaneously,the algorithm employs the technologies of random sampling,limiting the depth of the clustering feature tree and creating leaf node link to improve the temporal efficiency and the result of clustering.Experiments show that RSLDCH o...
出处
《郑州大学学报(理学版)》
CAS
2007年第3期80-83,共4页
Journal of Zhengzhou University:Natural Science Edition
关键词
层次聚类
BIRCH算法
CF
CFT
RSLDCH算法
hierarchical clustering
BIRCH algorithm
clustering feature
clustering feature tree
RSLDCH algorithm