摘要
针对密度峰值聚类(DPC)算法存在的dc值难选择及近邻原则聚合操作在低密度区效果不佳的问题,提出一种基于人工蜂群与CDbw聚类指标优化的密度峰值聚类(BeeDPC)算法,以实现类簇间数据点的自动识别和合理聚类,并解决DPC对类簇间数据点类别识别上存在的缺陷.实验结果表明,BeeDPC算法具有自动识别并合理聚类类簇间数据点、自动识别类簇中心点和类簇数量及自动处理任意分布数据集的优势.
Aiming at the problem that value of d c was difficult to select and the poor effect of neighborhood principle aggregation operation in low density area,we proposed a density peaks clustering(DPC)algorithmbased on artificial bee colony and CDbw clustering index optimization,which realized automatic identification and reasonable clustering of data points betweenclusters,and solved the defect of DPC in class identification of data points between clusters.Experiment results show that the BeeDPC algorithm has advantagesof automatic identification and reasonable clustering of data points between clusters,automatic identification of cluster centers and the number of clusters and dealing with arbitrary distributed data sets.
作者
姜建华
吴迪
郝德浩
王丽敏
张永刚
李克勤
JIANG Jianhua;WU Di;HAO Dehao;WANG Limin;ZHANG Yonggang;LI Keqin(Jilin Province Key Laboratory of Fintech,Department of Data Science, Jilin University of Finance and Economics,Changchun 130117,China;Symbol Computation and Knowledge Engineering of Ministry of Education,Jilin University,Changchun 130012,China;Department of Computer Science,State University of New York,New York 12561, USA)
出处
《吉林大学学报（理学版）》
CAS
CSCD
北大核心
2018年第6期1469-1475,共7页
Journal of Jilin University:Science Edition
基金
吉林省自然科学基金(批准号:20180101044JC)
吉林省社会科学规划基金(批准号:2018B79)
符号计算与知识工程教育部重点实验室项目和吉林财经大学科研项目(批准号:2018Z05).
关键词
聚类分析
CDbw评价指标
密度峰值
密度聚类
人工蜂群算法
cluster analysis
CDbw evaluation index
density peak
density clustering
artificial bee colony(ABC)algorithm
作者简介
姜建华(1979—),男,汉族,博士,副教授,从事数据挖掘、优化算法和电子商务的研究,E-mail:jjh@jlufe.edu.cn.;通信作者:李克勤(1963-),男,汉族,博士,教授,从事数据挖掘和高性能计算的研究,E-mail:lik@newpaltz.edu.
