北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2006, Vol. 29 ›› Issue (s2): 4-7.doi: 10.13190/jbupt.2006s2.4.294

• 论文 • 上一篇    下一篇

聚类有效性研究

任鸿1,郑岩2,曲昭伟3,邓芳2   

  1. 1.北京邮电大学 电子工程学院,北京100876; 2.北京邮电大学 智能通信软件与多媒体北京市重点实验室,北京100876;3.北京邮电大学 电子商务中心,北京100876
  • 收稿日期:2006-07-26 修回日期:1900-01-01 出版日期:2006-11-30 发布日期:2006-11-30
  • 通讯作者: 任鸿

Research on Clustering Validity

REN Hong 1,ZHENG Yan 2,QU Zhao-wei 3,DENG Fang 2   

  1. 1. School of Electronic Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China; 2. Beijing Key Laboratory of Intelligent Telecommunications Software and Multimedia, Beijing University of Posts and Telecommunications, Beijing 100876, China; 3. Research Center of Electronic Commerce, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2006-07-26 Revised:1900-01-01 Online:2006-11-30 Published:2006-11-30
  • Contact: REN Hong

摘要:

在对现有各种聚类有效性函数进行深入分析的基础上,提出了一种改进的聚类有效性函数,通过引入模拟退火机制,以获得最佳聚类结果。与其他经典的聚类有效性函数相比,改进后的聚类有效性函数一方面避免了单调性,另一方面又与数据集合本身的特征相关。仿真实验表明了其有效性。

关键词: 聚类有效性函数, 模糊c均值, 模拟退火

Abstract:

In order to obtain the optimal clustering result, an improved method is proposed by incorporating the simulated annealing mechanism based on analyzing the existing clustering validity functions. Comparing with the typical clustering validity functions, the advantages of this improved method will avoid the mono-tonicity and associating with the characteristics of the dataset. Experimental results show its effectiveness.

Key words: cluster validity function, fuzzy c means, simulated annealing

中图分类号: