Journal of Computer Applications ›› 2016, Vol. 36 ›› Issue (6): 1619-1623.DOI: 10.11772/j.issn.1001-9081.2016.06.1619

Previous Articles     Next Articles

Multi-view kernel K-means algorithm based on entropy weighting

QIU Baozhi1,2, HE Yanfang1, SHEN Xiangdong2   

  1. 1. School of Information Engineering, Zhengzhou University, Zhengzhou Henan 450001, China;
    2. Henan Provincial Institute of Scientific and Technical Information, Zhengzhou Henan 450003, China
  • Received:2015-11-10 Revised:2016-01-25 Online:2016-06-10 Published:2016-06-08
  • Supported by:
    This work is supported by the Basic and Advanced Technology Research Project of Henan Province (152300410191).

熵加权多视角核K-means算法

邱保志1,2, 贺艳芳1, 申向东2   

  1. 1. 郑州大学 信息工程学院, 郑州 450001;
    2. 河南省科学技术信息研究院, 郑州 450003
  • 通讯作者: 贺艳芳
  • 作者简介:邱保志(1964-),男,河南驻马店人,教授,博士,CCF会员,主要研究方向:数据挖掘、人工智能;贺艳芳(1988-),女,河南漯河人,硕士研究生,主要研究方向:数据挖掘;申向东(1966-),男,河南安阳人,高级工程师,主要研究方向:数据挖掘、科技统计。
  • 基金资助:
    河南省基础与前沿技术研究项目(152300410191)。

Abstract: In multi-view clustering based on view weighting, weight value of each view products great influence on clustering accuracy. Aiming at this problem, a multi-view clustering algorithm named Entropy Weighting Multi-view Kernel K-means (EWKKM) was proposed, which assigned a reasonable weight to each view so as to reduce the influence of noisy or irrelevant views, and then to improve clustering accuracy. In EWKKM, different views were firstly represented by kernel matrix and each view was assigned with one weight. Then, the weight of each view was calculated from the corresponding information entropy. Finally, the weight of each view was optimized according to the defined optimized objective function, then multi-view clustering was conducted by using the kernel K-means method.The experiments were done on the UCI datasets and a real datasets. The experimental results show that the proposed EWKKM is able to assign the optimal weight to each view, and achieve higher clustering accuracy and more stable clustering results than the existing cluster algorithms.

Key words: clustering, multi-view clustering, kernel K-means, entropy

摘要: 在基于视角加权的多视角聚类中,每个视角的权重取值对聚类结果的精度都有着重要的影响。针对此问题,提出熵加权多视角核K-means(EWKKM) 算法,通过给每个视角分配一个合理的权值来降低噪声视角或无关视角对多视角聚类的影响,进而提高聚类的精度。EWKKM算法中,首先用核矩阵表示不同的视角,给每个视角分配一个权重;然后,利用信息熵计算出各个视角的熵权重;最后,按照定义的目标函数对各个视角的权重进行优化,使用核K-means进行多视角聚类。在UCI数据集及人工数据集进行实验,实验结果表明熵加权多视角核K-means算法能够为每个视角分配一个最优的权重值,聚类的精确度优于已有的聚类算法,具有更稳定的聚类结果。

关键词: 聚类, 多视角聚类, K-means,

CLC Number: