• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2011, Vol. 33 ›› Issue (12): 121-125.

• 论文 • 上一篇    下一篇

模糊聚类分析系统的研究与实现

蒋德珑1,2,李〓盛3,王克文1,付金光1   

  1. (1.郑州大学电气工程学院,河南 郑州 450001;2.71781部队,河南 洛阳 471100;3.铁道警官高等专科学校,河南 郑州 450001)
  • 收稿日期:2011-04-11 修回日期:2011-09-28 出版日期:2011-12-24 发布日期:2011-12-25

Research and Implementation of a Fuzzy Clustering Analysis System

JIANG Delong1,2,LI Sheng3,WANG Kewen1,FU Jinguang1   

  1. (1.School of Electrical Engineering,Zhengzhou University,Zhengzhou 450001;2.Corps 71781,Luoyang 471100;3.Railway Police College,Zhengzhou 450001,China)
  • Received:2011-04-11 Revised:2011-09-28 Online:2011-12-24 Published:2011-12-25

摘要:

针对目前多种模糊聚类算法组合应用研究较少的现状,在分别对基于模糊等价关系和基于模糊等价划分这两种模糊聚类分析方法进行了较为深入研究的基础上,根据两种方法的特点,构建了一种组合式模糊聚类分析方法。采用VC++与Fortran6.5语言混合编程,开发了一套模糊聚类分析系统,分别实现了基于模糊等价关系法、基于模糊ISODATA法和基于组合式模糊聚类法的模糊聚类分析,并将其应用于电力系统中不良数据的辨识处理,取得了较为理想的效果。实例分析表明,组合式算法能够有效克服单独使用某种模糊聚类算法的缺点,同时也反映出该模糊聚类分析系统具有聚类方法选取灵活、软件界面友好、计算速度快等特点,有良好的应用前景。

关键词: 模糊聚类分析, 模糊等价关系, 模糊ISODATA法, 组合模型, 混合编程

Abstract:

By the analytic comparison between fuzzy equivalence and fuzzy ISODATA methods, two methods are synthesized to structure a combined fuzzy clustering analysis method. The set of fuzzy clustering analysis system realizes separately based on fuzzy equivalence relations, fuzzy equivalence relation and fuzzy clustering analysis based on integrated fuzzy clustering, which is developed with VC++ and Fortran6.5. The system is used to identify the baddata in power systems and has made an ideal effect. The analysis of example reveals that the system has the characteristics of flexible choice for clustering approaches, friendly interface, and fast speed of calculation, etc. Therefore, it will have a great application prospect in the future.

Key words: fuzzy clustering analysis;fuzzy equivalence relation;fuzzy ISODATA method;combined model;hybrid programming