【摘要】 BIRCH方法是一个集成的层次聚类方法.它克服了凝聚层次聚类方法所面临的两个难点:可伸缩性和不能撤销前一步工作的问题.基于BIRCH聚类的多阶段聚类算法思想,结合基于权重的欧式距离度量和基于划分的K-means算法,提出了一种基于加权相似性的BIRCH聚类方法,并将方法应用在时间序列的气象数据分析中.更多还原

【Abstract】 BIRCH is an integrated method of hierarchical clustering.It conquers two difficulties of condensation hierarchical clustering,which are the problems of scalability and rollback.Based on the idea of multi-stage clustering of the BIRCH algorithm,with the weight-based Euclidian distance metric and partition-based K-means algorithm,this paper proposes the BRICH clustering algorithm based on the weighted similarity,and the method is applied into the time series analysis of meteorology data.更多还原