一种带有显著特征项的模糊聚类算法

An Algorithm with Significant Feature for Fuzzy Clustering

模糊聚类分析可将特征相近的对象聚集为一类,但不直接显示聚类依据.改进算法根据对象特征项对聚类分析的贡献率,提出"显著特征项"的概念,进而得出聚类的主要依据.仿真结果表明,算法的聚类依据明确,数学基础可靠,计算过程规范.

Fuzzy clustering analysis aims to integrate the object ingredients of similar features without demonstrating the evidence for clustering.This paper gives a notion of what is called ＂significant feature＂ in improving the method according to the contribution rate of the object features for clustering,and the main evidence of clustering can be figured out.Result of simulation proves that the evidence of clustering is valid and of practical use due to its reliability in computing method and the regularity in the process of computation.