运用区间二型模糊集的半监督模糊聚类算法  

Partial Supervised Fuzzy Clustering Algorithm Based onInterval Type-2 Fuzzy Sets

作　　者：杨昔阳[1,2,3] 陈豪 张诗晴 YANG Xiyang;CHEN Hao;ZHANG Siqing(Fujian Provincial Key Laboratory of Data-Intensive Computing(Quanzhou Normal University),Quanzhou Fujian 362000,China;Key Laboratory of Intelligent Computing and Information Processing(Quanzhou Normal University),Quanzhou Fujian 362000,China;Fujian Key Laboratory of Financial Information Processing(Putian University),Putian Fujian 351100,China)

机构地区：[1]福建省大数据管理新技术与知识工程重点实验室(泉州师范学院),福建泉州362000 [2]智能计算与信息处理福建省高等学校重点实验室(泉州师范学院),福建泉州362000 [3]福建省金融信息处理重点实验室(莆田学院),福建莆田351100

出　　处：《泉州师范学院学报》2023年第2期1-8,共8页Journal of Quanzhou Normal University

基　　金：福建省自然科学基金(2021J01001);泉州市科技局资助项目(2021N042);福建省金融信息处理重点实验室(莆田学院)开放课题(JXC202205);国家大学生创新创业训练项目(202210399013)。

摘　　要：将区间二型模糊集引入半监督聚类,提出一种运用区间二型模糊集的半监督聚类算法.该算法以区间二型模糊集为隶属度,将已标识样本的类别标签融入目标函数,使得聚类算法可以在已标识样本类别信息的引导下,得到合理的隶属度.实验表明,相比其他聚类模型,结合半监督机制和区间二型模糊集的聚类算法具有更好的聚类性能.In this paper,interval type-2 fuzzy sets are introduced into partial supervised clustering,and a partial supervised clustering algorithm based on interval type-2 fuzzy sets is proposed.This algorithm takes interval type-2 fuzzy sets as the membership degree,and integrates the class label of the labeled samples into the objective function,so that the clustering algorithm can get a reasonable membership degree under the guidance of the label information of the labeled samples.Experiments show that the clustering algorithm combining partial supervised mechanism and interval type-2 fuzzy sets has better clustering performance than other clustering models.

关 键 词：半监督聚类算法 区间二型模糊集 模糊聚类 半监督机制 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]