基于半监督判别最大熵模糊聚类算法  

Semi-supervised Discriminative Fuzzy Maximum Entropy Clustering Algorithm with Pairwise Constraints

作　　者：麦晓冬[1] MAI Xiao-dong(College of Information Technology, Guangdong Industry Polytechnic, Guangzhou 510300, China)

机构地区：[1]广东轻工职业技术学院信息技术学院,广东广州510300

出　　处：《内蒙古师范大学学报（自然科学汉文版）》2017年第5期763-766,共4页Journal of Inner Mongolia Normal University(Natural Science Edition)

基　　金：广东省高等学校优秀青年教师培养计划(YQ2015172)

摘　　要：为了解决大量高维数据分类的问题,给出一种基于半监督判别最大熵模糊的聚类算法.该算法不仅继承了已有FLDA-MEFCA算法的降维优势,而且可以充分利用监督信息来提高聚类性能.实验证明该算法的总体性能优于最大熵模糊聚类算法、FLDA-MEFCA和经典FCM类算法.In order to solve the problem of classifying large number of multi-dimensional data,one semi-supervised discriminative fuzzy maximum entropy clustering algorithm with pairwise constraints is proposed in this paper.This algorithm not only inherits the advantages of the dimension reduction of FLDA-MEFCA,but also can take advantage of supervised information to improve the performance of clustering.Experiments show that the overall performance of the proposed algorithm is superior to Maximum Entropy Fuzzy Clustering Algorithm（MEFCA）,FLDA-MEFCA and classical FCM clustering algorithm.

关 键 词：模糊聚类 降维 判别聚类 半监督聚类 成对约束 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]