摘要
针对二分K-means算法存在的误判实例无法再参与后续划分并降低了聚类的精度的问题.提出一种基于部分实例重判的二分K-means算法,通过区分目标簇和候选簇,过滤出候选簇中的召回实例,对召回实例所应归属的簇进行重判,实现了误判实例的正确聚类.实验结果表明,改进算法对三个实验数据集都是有效的,在不同程度上提高了聚类的准确性,同时对算法的运行速度也有小幅度的提升.
The problem of misjudgment instance of bisecting K-means being unable to participate in the subsequent partitioning reduces the accuracy of clustering. This paper proposes a bisecting K-means algorithm based on partial instance rejudge,which can correctly classify the misjudgment instances by distinguishing the object clusters and the candidate clusters,filtering the recall instances from the candidate clusters,and reclassifying the recall instances. The experimental results show that the improved algorithm is effective for three data sets,and can improve the accuracy of clustering in different extent and the running speed of the algorithm.
作者
吴清寿
刘耿耿
郭文忠
WU Qingshou;LIU Genggeng;GUO Wenzhong(Department of Mathematics and Computer Science,Wuyi University, Wuyishan,Fujian 354300,China;College of Mathematics and Computer Science,Fuzhou University, Fuzhou,Fujian 350116,China)
出处
《福州大学学报（自然科学版）》
CAS
北大核心
2018年第3期317-323,共7页
Journal of Fuzhou University(Natural Science Edition)
基金
国家自然科学基金资助项目(11501114)
福建省教育厅科技资助项目(JA14309)
福建省中青年教师教育科研资助项目(JAT170608)
关键词
二分k均值
部分实例重判
候选簇
召回实例
聚类
bisecting k-means
partial instance rejudge
candidate cluster
recall instance
clustering
