计算机应用 ›› 2005, Vol. 25 ›› Issue (06): 1357-1359.DOI: 10.3724/SP.J.1087.2005.1357

• 数据库与数据挖掘 • 上一篇    下一篇

基于关联函数的动态聚类算法及应用

朱群雄,宣达婧,顾祥柏   

  1. 1.北京化工大学信息科学与技术学院; 2.中国石化工程建设公司
  • 发布日期:2011-04-06 出版日期:2005-06-01
  • 基金资助:

    教育部科学技术研究重点项目(01024);;中石化科学技术研究开发项目(E03007)

Dynamic clustering algorithm based on dependent function and its application

ZHU Qun-xiong1, XUAN Da-jing1, GU Xiang-bai1,2   

  1. 1. School of Information Science & Technology, Beijing University of Chemical Technology, Beijing 100029,China; 2. Sinopec Engineering Inc, Beijing 100101,China
  • Online:2011-04-06 Published:2005-06-01

摘要: 根据时序立体数据的特点,提出了基于关联函数一致性矩阵的动态聚类算法。给出了适用于时序立体数据关联函数的改进标准关联函数计算公式,并将该算法应用于乙烯裂解炉报警系统,结合流程的时序立体数据,得到了裂解炉报警系统的动态聚类分类结果,并验证了提出算法的有效性。本文算法对于时序数据的聚类具有较强的鲁棒性。

关键词: 动态聚类, 关联函数, 一致性矩阵, 乙烯裂解炉

Abstract: A dynamic clustering algorithm was proposed based on consistent matrix of dependent function for time series multi-dimensional data. Further, improved standard dependent function calculation formula was proposed for time series cubic data. The provided algorithm was adopted to cluster ethylene cracking furnace alarm information by time series data. The effective of the proposal algorithm has been verified by result of above case study. Proposed dynamic clustering algorithm has strong robustness in clustering of time series multi-dimensional data.

Key words: dynamic clustering algorithm, dependent function, consistent matrix, ethylene cracking furnace

中图分类号: