Journal of Computer Applications ›› 2019, Vol. 39 ›› Issue (3): 742-749.DOI: 10.11772/j.issn.1001-9081.2018071617

Previous Articles     Next Articles

Sparse non-negative matrix factorization based on kernel and hypergraph regularization

YU Jianglan1,2, LI Xiangli1,2,3, ZHAO Pengfei4   

  1. 1. School of Mathematics and Computing Science, Guilin University of Electronic Technology, Guilin Guangxi 541004, China;
    2. Guangxi Key Laboratory of Cryptography and Information Security(Guilin University of Electronic Technology), Guilin Guangxi 541004, China;
    3. Guangxi Colleges and Universities Key Laboratory of Data Analysis and Computation(Guilin University of Electronic Technology), Guilin Guangxi 541004, China;
    4. School of Mathematics, Southwest Jiaotong University, Chengdu Sichuan 611756, China
  • Received:2018-08-03 Revised:2018-09-22 Online:2019-03-10 Published:2019-03-11
  • Contact: 李向利
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (11601012, 71561008), the Guangxi Key Laboratory of Cryptography and Information Security (GCIS201708), the Guangxi Natural Science Foundation (2018GXNSFAA138169), the Postgraduate Excellent Dissertation Cultivation Project of Guilin University of Electronic Technology (16YJPYSS22).

基于核技巧和超图正则的稀疏非负矩阵分解

余江兰1,2, 李向利1,2,3, 赵朋飞4   

  1. 1. 桂林电子科技大学 数学与计算科学学院, 广西 桂林 541004;
    2. 广西密码学与信息安全重点实验室(桂林电子科技大学), 广西 桂林 541004;
    3. 广西高校数据分析与计算重点实验室(桂林电子科技大学), 广西 桂林 541004;
    4. 西南交通大学 数学学院, 成都 611756
  • 作者简介:余江兰(1993-),女,重庆人,硕士研究生,主要研究方向:最优化理论及其应用;李向利(1977-),女,陕西宝鸡人,副教授,博士,主要研究方向:最优化理论及其应用;赵朋飞(1993-),男,陕西咸阳人,硕士研究生,主要研究方向:大数据与数据科学。
  • 基金资助:
    国家自然科学基金资助项目(11601012,71561008);广西密码学与信息安全重点实验室研究课题(GCIS201708);广西自然科学基金资助项目(2018GXNSFAA138169);桂林电子科技大学研究生优秀学位论文培育项目(16YJPYSS22)。

Abstract: Focused on the problem that when traditional Non-negative Matrix Factorization (NMF) is applied to clustering, robustness and sparsity are not considered at the same time, which leads to low clustering performance, a sparse Non-negative Matrix Factorization algorithm based on Kernel technique and HyperGraph regularization (KHGNMF) was proposed. Firstly, on the basis of inheriting good performance of kernel technique, L2,1 norm was used to improve F-norm of standard NMF, and hyper-graph regularization terms were added to preserve inherent geometric structure information among the original data as much as possible. Secondly, L2,1/2 pseudo norm and L1/2 regularization terms were merged into NMF model as sparse constraints. Finally, a new algorithm was proposed and applied to image clustering. The experimental results on six standard datasets show that KHGNMF can improve clustering performance (accuracy and normalized mutual information) by 39% to 54% compared with nonlinear orthogonal graph regularized non-negative matrix factorization, and the sparsity and robustness of the proposed algorithm are increased and the clustering effect is improved.

Key words: Non-negative Matrix Factorization (NMF), hypergraph regularization, L2,1/2-matrix pseudo norm, sparsity, robustness, L2,1-norm

摘要: 针对传统的非负矩阵分解(NMF)应用于聚类时,没有同时考虑到鲁棒性和稀疏性,导致聚类性能较低的问题,提出了基于核技巧和超图正则的稀疏非负矩阵分解算法(KHGNMF)。首先,在继承核技巧的良好性能的基础上,用L2,1范数改进标准非负矩阵分解中的F范数,并添加超图正则项以尽可能多地保留原始数据间的内在几何结构信息;其次,引入L2,1/2伪范数和L1/2正则项作为稀疏约束合并到NMF模型中;最后,提出新算法并将新算法应用于图像聚类。在6个标准的数据集上进行验证,实验结果表明,相对于非线性正交图正则非负矩阵分解方法,KHGNMF使聚类性能(精度和归一化互信息)成功地提升了39%~54%,有效地改善和提高了算法的稀疏性和鲁棒性,聚类效果更好。

关键词: 非负矩阵分解, 超图正则, L2,1/2矩阵伪范数, 稀疏性, 鲁棒性, L2,1范数

CLC Number: