摘要
在模糊C均值算法的基础上,通过对原有算法进行改进,以达到加快聚类速度的目的。提出了一种使用最速下降法来优化模糊C均值算法的方法。从传统的模糊C均值算法中推导出关于聚类中心的泛函迭代序列,并证明了该序列的收敛性,以及该序列收敛到的不动点是目标函数达到的极值点。而后,使用最速下降法加快该序列收敛速度。最终通过实验结果来验证了理论的可行性。在其迭代过程中,对于越偏离理论聚类中心的点,下降趋势比传统模糊C聚类算法就越明显。
In order to accelerate the rate of clustering, the fuzzy c-means algorithm has been improved. The method of using steepest descent method to optimize the fuzzy c-means algorithm is proposed. A functional iterative sequence about clustering centers is derived from the classical fuzzy c-means algorithm, the convergence of the squence is also proved. It is shown that the fixed point of the squence is an extremum of the objective function. Then, the steepest descent method is used to accelerate the convergence of the sequence. The experimental results show that the theory is right, the further of the distance between the point and real center, the greater of the rate of convergence than the traditional fuzzy c-means algorithm.
出处
《科学技术与工程》
北大核心
2012年第30期7915-7919,共5页
Science Technology and Engineering
关键词
聚类
模糊C均值
泛函迭代序列
最速下降法
clustering ,fuzzy c-means algorithm, functional iterative sequence, the steepest descent method