Journal of Computer Applications ›› 2017, Vol. 37 ›› Issue (4): 1071-1074.DOI: 10.11772/j.issn.1001-9081.2017.04.1071

Previous Articles     Next Articles

Incremental learning algorithm based on graph regularized non-negative matrix factorization with sparseness constraints

WANG Jintao, CAO Yudong, SUN Fuming   

  1. School of Electronics and Information Engineering, Liaoning University of Technology, Jinzhou Liaoning 121001, China
  • Received:2016-08-09 Revised:2016-10-14 Online:2017-04-10 Published:2017-04-19
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (61572244), the Program for Liaoning Excelent Talents in Universities (LR2015030).

稀疏约束图正则非负矩阵分解的增量学习算法

汪金涛, 曹玉东, 孙福明   

  1. 辽宁工业大学 电子与信息工程学院, 辽宁 锦州 121001
  • 通讯作者: 曹玉东
  • 作者简介:汪金涛(1992-),男,安徽合肥人,硕士研究生,主要研究方向:模式识别;曹玉东(1971-),男,辽宁铁岭人,副教授,博士,主要研究方向:图像处理、模式识别;孙福明(1972-),男,辽宁大连人,教授,博士,CCF会员,主要研究方向:机器学习、模式识别。
  • 基金资助:
    国家自然科学基金资助项目(61572244);辽宁省高等学校优秀人才支持计划项目(LR2015030)。

Abstract: Focusing on the issues that the sparseness of the data obtained after Non-negative Matrix Factorization (NMF) is reduced and the computing scale increases rapidly with the increasing of training samples, an incremental learning algorithm based on graph regularized non-negative matrix factorization with sparseness constraints was proposed. It not only considered the geometric structure in the data representation, but also introduced sparseness constraints to coefficient matrix and combined them with incremental learning. Using the results of previous factorization involved in iterative computation with sparseness constraints and graph regularization, the cost of the computation was reduced and the sparseness of data after factorization was highly improved. Experiments on both ORL and PIE face recognition databases demonstrate the effectiveness of the proposed method.

Key words: Non-negative Matrix Factorization (NMF), sparse constraint, graph regularization, geometry, incremental learning

摘要: 针对非负矩阵分解后数据的稀疏性降低、训练样本增多导致运算规模不断增大的现象,提出了一种稀疏约束图正则非负矩阵分解的增量学习算法。该方法不仅考虑数据的几何信息,而且对系数矩阵进行稀疏约束,并将它们与增量学习相结合。算法在稀疏约束和图正则化的条件下利用上一步的分解结果参与迭代运算,在节省大量运算时间的同时提高了分解后数据的稀疏性。在ORL和PIE人脸数据库上的实验结果表明了该算法的有效性。

关键词: 非负矩阵分解, 稀疏约束, 图正则, 几何结构, 增量学习

CLC Number: