一种FCM聚类算法的改进与优化

针对一种FCM聚类算法的一些不足 ,提出了具体的改进与提高的方法 ,并引入模糊决策理论 ,进行算法参数m的优化选择。改进后的算法 ,一方面有效地弥补了原算法中存在的不足 ,更好地解决了聚类数目选择等初值问题 ;另一方面 ,通过对参数m的优化选择 ,取得了较理想的聚类效果。最后给出了几种聚类算法对某数据样本集的聚类对比结果。

Improvement and Optimization of a Fuzzy C-Means Clustering Algorithm

In this paper, an approach is proposed to improve a FCM clustering algorithm and the fuzzy decision theory is adopted to optimize the parameter m in the algorithm. On one hand, the proposed algorithm resolves the problem of the initial value effectively; and on the other hand, the better clustering results are obtained through the optimal choice of the parameter m . Finally, a few results are demonstrated to illustrate the effectiveness of the proposed method.