中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊

Journal of Chongqing Jiaotong University(Natural Science) ›› 2013, Vol. 32 ›› Issue (4): 725-728.DOI: 10.3969 /j.issn.1674-0696.2013.04.41

Previous Articles    

Clustering Algorithm for High Dimensional Data Stream Based on Projection and Density

Wang Renhong,Wang Jiawei,Liang Zongbao   

  1. School of Information Science & Engineering,Chongqing Jiaotong University,Chongqing 400074,China
  • Received:2012-05-27 Revised:2013-04-23 Online:2013-08-15 Published:2014-10-27

基于投影和密度的高维数据流聚类算法

汪仁红,王家伟,梁宗保   

  1. 重庆交通大学 信息科学与工程学院,重庆 400074
  • 作者简介:汪仁红(1987—),女,重庆人,硕士研究生,主要从事数据流处理方面的研究。E-mail:wangrenhong020811112@126.com。
  • 基金资助:
    交通运输部西部项目( 20113188141480)

Abstract: A clustering algorithm for high dimensional data stream based on the projection and density ( HpDenStream) is proposed,which is based on the classical data stream clustering algorithm. This algorithm combines with the sliding window technical,and it uses the projection clustering algorithm to reduce the data dimensionality of dimensional data streams and then adopts the density clustering algorithm to detect the anomaly data detection. The simulation results show that the algorithm not only takes up smaller storage space and less workload,but also improves the efficiency of the implementation of the algorithm.

Key words: data stream, clustering algorithm, projection, dimension reduction, density, anomaly detection

摘要: 在经典数据流的聚类算法基础之上,提出了一种基于投影和密度的高维数据流聚类算法———HpDenStream,该算法结合滑动窗口技术,采用投影算法对高维数据流进行降维处理,并运用密度聚类算法对降维后的数据进行异常数据检测。仿真实验结果表明: 该方法占用的存储空间小,算法的工作量少,并提高了算法的执行效率。

关键词: 数据流, 聚类算法, 投影, 降维, 密度, 异常检测

CLC Number: