计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (5): 206-209.

• 图形、图像、模式识别 • 上一篇    下一篇

一个改进的模糊聚类有效性指标

朱文婕1,2,吴 楠1,胡学钢1   

  1. 1.合肥工业大学 计算机与信息学院,合肥 230009
    2.蚌埠医学院,安徽 蚌埠 233030
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-02-11 发布日期:2011-02-11

Improved cluster validity index for fuzzy clustering

ZHU Wenjie1,2,WU Nan1,HU Xuegang1   

  1. 1.School of Computer & Information,Hefei University of Technology,Hefei 230009,China
    2.Bengbu Medical College,Bengbu,Anhui 233030,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-02-11 Published:2011-02-11

摘要: 聚类有效性指标既可用来评价聚类结果的有效性,也可以用来确定最佳聚类数。根据模糊聚类的基本特性,提出了一种新的模糊聚类有效性指标。该指标结合了数据集的分布特征和数据隶属度两个重要因素来评价聚类结果,提高了判别的准确性。实验证明,该指标能对模糊聚类结果进行正确的评价,并自动获得最佳聚类数,特别是对类间有交叠的情况能够做出准确判定。

关键词: 有效性指标, 模糊聚类分析, 模糊C均值聚类(FCM)算法

Abstract: Cluster validity index is used to scale the validity of clustering,and determine the number of clusters.According to the basic properties of fuzzy clustering,a new validity index for fuzzy clustering is proposed.It exploits two important evaluation factors:Dataset distribution and fuzzy membership grade.The experiment results indicate that the index is effective.And it allows the number of clusters to be determined automatically.Especially,the new index can correctly identify the optimal clustering number under the condition of overlapping clusters.

Key words: cluster validity index, fuzzy clustering, Fuzzy C-Means(FCM)