摘要
为了解决模糊C-均值(FCM)聚类算法在使用欧氏距离计算样本与类中心点的距离时计算量大的问题,提出了一种基于属性约简的FCM聚类算法。该算法根据粗糙集理论对初始数据进行属性约简,消除数据对象中的冗余值,然后再对约简后的属性集进行模糊聚类。实验结果表明,该算法能有效减少FCM算法的距离函数计算量,在不降低聚类精度的前提下,提高了FCM算法的执行效率。
To solve the problem that amount of computation is too large when the distance between the multi-attribute data sample and the center of the class is calculated using Euclidean distance function in fuzzy C-means （FCM） clustering algorithm, an FCM clustering algorithm based on attribute reduction is proposed. In the proposed algorithm, first, attributes of the initial data are reduced based on the rough set theory, and the redundant values of data objects are eliminated; and then fuzzy clustering is carried out on the reduction attribute sets. Experimental results show that the amount of computation of the distance function of FCM algorithm is effectively reduced in the algorithm, and the efficiency of the implementation of FCM algorithm is improved without reducing the prosion of clustering.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第18期4062-4064,4127,共4页
Computer Engineering and Design
基金
国家火炬计划基金项目(2004EB33006)
江苏省高校自然科学指导性计划基金项目(05JKD520050)
关键词
模糊划分
FCM聚类
粗糙集
属性约简
区分矩阵
fuzzy partition
FCM algorithm
rough set
attribute reduction
discernibility matrix
