基于模糊稀疏度的聚类分析

Sparseness-Based Clustering Techniques

本文简要评价了模糊聚类分析中常用的两种方法:λ截集聚类法和模糊ISOD-ATA 聚类法,指出了它们在应用方面的一些局限性,据此提出了一种基于模糊稀疏度的聚类技术——SBCT 法。其中,改进了现有贴近度的概念,提出了用于衡量模糊幂集中模糊子集分布状况的稀疏度概念。

This paper gives an inquiry on the fuzzy clustering problem.It discusses both the fuzzy λ-clustering method and the fuzzy ISODATA clustering method,and finds some conclusions about their problems in practice.To overcome the shortcomings of the two techniques,the concept of sparseness based concept of fuzzy nearness and Sparseness-Based Clustering Technique(SBCT) are presented.