基于区间二型模糊集的模糊等价关系聚类分析

发布时间：2018-11-20 06:48

【摘要】：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.
[Abstract]:The traditional clustering algorithm has great limitation in solving the clustering problem with uncertainty. In order to solve the uncertainty of clustering problem better, this paper is based on interval type 2 fuzzy set theory. A clustering analysis algorithm based on type 2 fuzzy equivalence relation is proposed. Firstly, the information of language variables is completely transformed into interval type 2 fuzzy sets, and then the advantages of language variables and interval type 2 fuzzy sets are combined, and the Jaccard similarity of interval type 2 fuzzy sets is obtained. A new fuzzy equivalence relation clustering analysis method based on interval type 2 fuzzy language variables is proposed, and a specific algorithm flow is designed. Compared with the traditional fuzzy equivalence relation clustering algorithm, the new clustering algorithm has better ability to deal with uncertain problems and avoids the loss of information in the process of clustering calculation. At the same time, the new clustering algorithm can give the dynamic clustering results with the variation of clustering similarity parameters. Finally, the paper takes the mobile phone brand clustering based on ecommerce platform as an example to verify the feasibility and rationality of the new algorithm.
【作者单位】：东南大学经济管理学院;南京农业大学金融学院;
【基金】：国家自然科学基金(71171048,71371049,71501098) 教育部博士点基金(20120092110038)~~
【分类号】：O159

【相似文献】