一种含噪数据FCM聚类算法

FCM Clustering Algorithm for Noisy Data

针对IADFCM算法在运算过程中忽略区间中点和半宽对区间数分析的问题,给出基于中点、半宽含权重区间数间的欧氏距离,提出一种改进的聚类分析算法,对模拟数据集和实际数据集分别进行仿真实验,实验结果表明,该算法是有效的。

Aiming at the problems that it ignores the different effect which the median and wide of interval play on the analysis of interval number, this paper presents a new measure of distance based on the median and wide of interval with different weight. On basis of this, a revised clustering algorithm is proposed. The experiments of the simulative and practical datasets are made by this algorithm and IADFCM. Experimental results indicate this algorithm is effective.