一种新型SFCM聚类算法

A new algorithm for Semi-Fuzzy c-Means clustering

为提高故障诊断模式分类的收敛速度及诊断准确度,采用阈值化类内距离的方法,研究了一种新型SFCM聚类算法,证明了算法模糊加权幂指数m在区间(0,1)取值时能实现半模糊聚类,讨论了阈值η对算法的影响并给出了聚类程序算法,实验结果表明此算法较FCM算法在收敛速度和聚类精度方面有较好表现,在故障模式分类中有一定的使用价值。

In order to improve the convergence speed and precision in pattern classification of fault diagnosis, a Semi-Fuzzy c Means（SFCM） algorithm was proposed based on revised Euclidean distance. We proved that Se- mi-Fuzzy clustering can be realized when the fuzzy weighted exponent sign m is defined in （0,1）. We also dis- cussed the influence of threshold parameter η on clustering algorithm, and presented the program steps. The experimental results showed that the proposed algorithm is better than Fuzzy c-Means （FCM） clustering algorithm in convergence speed and clustering precision, and thus can be used in fault pattern classification.