基于主元的多元时间序列聚类分析方法研究
首发时间:2012-06-25
摘要:为提高多元时间序列聚类算法的效率,采用基于主元分析的多元时间序列聚类方法. 将原始MTS数据线性组合形成一系列互不相关的簇,利用每个簇的代表元素与剩余元素的前 个主元与之间的扩展Euclid范数对选取的多元时间序列的主元进行聚类分析。理论分析和实验结果表明该算法聚类质量结果和运行时间明显优于直接利用k-均值法时的聚类结果。
For information in English, please click here
analysis method research of multivariate time series clustering based on principal comonent
Abstract:To improve the efficiency of algorithms for Multivariate Time Series data clustering, an efficient clustering algorithm for Multivariate Time Series-PCA-CLUSTER is proposed. A series of unrelated clusters are obtained by linear combination, based on Eros distance between residual MTS data and cluster representing elements, the original MTS data dimension is reduced. Subsequently, K-neighbor clustering analysis was carried out on the principal component series of MTS, and finally K MTS cluster are obtained. The experimental results show that, the clustering quality and operation time of proposed algorithm are better than that of direct K-mean clustering method obviously.
Keywords: Multivariate Time Series Principal Component Analysis Clustering Analysis
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于主元的多元时间序列聚类分析方法研究
评论
全部评论0/1000