模糊c均值聚类算法中参数m的优选 |
| |
引用本文: | 高新波,李洁,谢维信.模糊c均值聚类算法中参数m的优选[J].模式识别与人工智能,2000,13(1). |
| |
作者姓名: | 高新波 李洁 谢维信 |
| |
作者单位: | 1. 西安电子科技大学电子工程学院西安 710071 2. 深圳大学深圳 518060 |
| |
摘 要: | 本文利用模糊决策理论提出了一种模糊c均值(FCM)聚类算法中加权指数m的优选方法.文中定义了合适的模糊目标和模糊约束,通过模糊决策确定最佳的m值,以保证FCM算法获得好的聚类效果.实验结果显示了该方法的有效性,并得到实际应用中m的最佳取值范围为1.5,2.5].
|
关 键 词: | 模糊聚类 加权指数 模式识别 模糊决策 |
OPTIMAL CHOICE OF WEIGHTING EXPONENT IN A FUZZY C-MEANS CLUSTERING ALGORITHM |
| |
Abstract: | This paper proposes a technique for determining the optimal value of the weighting exponent, m, a parameter in a fuzzy c-means (FCM) algorithm, using the concept of fuzzy decision theory. A proper fuzzy goal and fuzzy constraint being defined, the optimal choice of m is made by fuzzy decision, such that given m , the FCM algorithm produces “good” clusters. Experimental results demonstrate its effectiveness and arrive at a conclusion that the optimal selected range of m is 1.5, 2.5]. |
| |
Keywords: | Fuzzy Clustering Weighting Exponent Pattern Recognition Fuzzy Decision |
本文献已被 万方数据 等数据库收录! |
|