基于模糊测度和证据理论的模糊聚类集成方法
CSTR:
作者:
作者单位:

空军工程大学防空反导学院,西安710051.

作者简介:

毕凯

通讯作者:

中图分类号:

TP391

基金项目:

国家自然科学基金项目(60975026, 61273275).


Fuzzy clustering ensemble based on fuzzy measure and DS evidence theory
Author:
Affiliation:

School of Air and Missile Defense,Air Force Engineering University,Xi’an 710051,China.

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    针对现有集成方法在处理模糊聚类时存在的不足, 提出一种基于证据理论的模糊聚类集成方法. 以各聚类成员作为证据元, 以样本点间的类别关系作为焦元, 通过证据积累构造互相关矩阵. 考虑到模糊聚类对于各样本点的聚类有效性, 提出一种结合点模糊度和模糊贴近度的类别关系表示方法, 并以此作为各证据元的基本概率赋值函数. 最后基于互相关矩阵构造样本点间相似性关系, 并利用谱聚类算法对其聚类. 实验中通过与多种已有聚类集成方法的对比表明, 该方法具有较高的聚类性能.

    Abstract:

    In order to solve the weakness of present ensemble methods for fuzzy clustering, a method of fuzzy clustering ensemble based on Dempster-Shafer(DS) theory is proposed, which takes every cluster member as the evidence, takes the relationship between pair of dates as the event and makes the co-matrix by accumulation of the evidence. Considering the classification effectiveness of each data point, a relationship based on the fuzzy degree and the fuzzy close-degree is proposed, and the new relationship is considered as the basic probability assignment for the evidence. Finally, the spectral clustering algorithm is used to make the final partition by similarity relationship based on the co-association matrix. Experimental results show that the method proposed is better than the current methods for clustering ensemble.

    参考文献
    相似文献
    引证文献
引用本文

毕凯 王晓丹 邢雅琼.基于模糊测度和证据理论的模糊聚类集成方法[J].控制与决策,2015,30(5):823-830

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2014-03-16
  • 最后修改日期:2014-07-16
  • 录用日期:
  • 在线发布日期: 2015-05-20
  • 出版日期:
文章二维码