基于区间二型模糊集的模糊等价关系聚类分析

传统聚类算法在解决含有不确定性的聚类问题时具有很大的局限性,为了更好地解决聚类问题中的不确定性,论文基于区间二型模糊集理论,提出了基于二型模糊等价关系的聚类分析算法.论文首先将语言变量信息完整地转化为区间二型模糊集,接着把语言变量和区间二型模糊集的优势相结合,通过区间二型模糊集的Jaccard相似度,提出了基于区间二型模糊语言变量的模糊等价关系聚类分析新方法,并设计了具体的算法流程.新聚类算法相对于传统的模糊等价关系的聚类算法,具有更好地处理不确定性问题的能力,避免了聚类计算过程中的信息丢失.同时新聚类算法可以灵活给出随聚类相似性参数变化的动态聚类结果.论文最后以电商平台的手机品牌聚类为例,验证了新算法的可行性和合理性.

Clustering analysis of fuzzy equivalence based on interval type-2 fuzzy sets

Traditional clustering methods have great limitations when they deal with the clustering problems that involved uncertainties. In order to deal with the uncertainties in clustering analysis, this paper proposes a new clustering analysis approach based on type-2 fuzzy equivalence. It first transfers the linguistic variables into interval type-2 fuzzy sets (IT2FSs), and combines advantages of the linguistic variables and IT2FSs together. Then with the aid of the Jaccard similarity method, a new fuzzy equivalence clustering analysis method with specific algorithm processes based on IT2FSs is proposed. The new method can avoid the information loss in the process of clustering computations. Furthermore, the new method can produce the dynamic clustering results with the change of cluster similarity parameters in a flexible way. Finally, an example of the clustering on mobile phone brands under e-commerce platform is given to demonstrate the feasibility and rationality of the new method.