Abstract: Synthesizing and analyzing the standard Fuzzy C-Means（FCM） clustering algorithm and conditional FCM clustering algorithm, the fuzzy partition space is modified, and the constraints of fuzzy partition matrix are weakened, this paper proposes an extended conditional FCM clustering algorithm. The partition matrix and prototypes of the proposed algorithm do not rely on the context constraints and the total membership of fuzzy partition matrix. Experimental results show that the proposed algorithm can produce different clustering prototypes, and has excellent clustering performance.

Key words: Fuzzy C-Means（FCM） clustering, Conditional Fuzzy C-Means(CFCM) clustering, fuzzy partition matrix

摘要： 在综合分析标准的模糊C-均值聚类算法和条件模糊C-均值聚类算法基础上，对模糊划分空间进行修改，进一步弱化模糊划分矩阵的约束，给出一种扩展的条件模糊C-均值聚类算法。算法的划分矩阵和原型不依赖于背景约束及模糊划分矩阵的隶属度总和。实验结果表明：该算法可以得到不同的聚类原型，并具有很好的聚类效果。

关键词: 模糊C-均值聚类（FCM）, 条件模糊C-均值聚类（CFCM）, 模糊划分矩阵