Journal of Systems Engineering and Electronics ›› 2009, Vol. 31 ›› Issue (11): 2732-2735 .

• 软件、算法与仿真 • 上一篇    下一篇

基于目标函数的直觉模糊集合数据的聚类方法

申晓勇1, 雷英杰1,李 进1, 蔡 茹1,2   

  1. 1. 空军工程大学导弹学院, 陕西 三原 713800 2. 中国人民解放军94994部队, 江苏 南京 210036
  • 出版日期:2009-11-26 发布日期:2010-01-03

Clustering technique to intuitionistic fuzzy sets data based on objective function

SHEN Xiaoyong,LEI Yingjie,LI Jin,CAI Ru   

  1. 1. Missile Inst., Air Force Engineering Univ., Sanyuan 713800, China; 2. Unit 94994 of the PLA, Nanjing 210036, China
  • Online:2009-11-26 Published:2010-01-03

摘要:

针对直觉模糊集合数据的聚类问题,提出了一种基于目标函数的聚类方法。该方法定义了直觉模糊集合间的加权相似性准则,解决了数据聚类过程中各维特征分配不均匀的问题。通过增加非隶属度参数对模糊c〖CD*2〗均值(fuzzy cmeans, FCM)聚类算法中的模糊划分矩阵〖WTHX〗U〖WTBZ〗和目标函数进行改造,进而给出迭代推导公式和算法描述,把聚类归结为一个带约束的线性规划问题,适用于大数据量的情况。最后通过典型实例验证了该方法的有效性和优越性。

Abstract:

Aiming at the fuzzy clustering problems of intuitionistic fuzzy sets data, a clustering technique based on objective function is proposed. For solving the problem of every character with asymmetrical weight, the method defines the weighted similar principle among intuitionistic fuzzy sets. By adding the nonmembership degree parameter to the partition matrix 〖WTHX〗U〖WTBZ〗 and objective function, the fuzzy cmeans (FCM) clustering algorithm is improved. Furthermore, the iterative reasoning formula and the algorithm description are presented. Thus the technique ascribes the clustering to a linear restricted programming issue, which is applied to the status of vast samples. Finally, the validity and superiority of the proposed technique are checked with an classical instance.