| 稀疏约束图正则非负矩阵分解 |
| |
| 引用本文： | 姜伟,李宏,余霞国,杨炳儒.稀疏约束图正则非负矩阵分解[J].计算机科学,2013,40(1):218-220,256. |
| |
| 作者姓名： | 姜伟 李宏 余霞国 杨炳儒 |
| |
| 作者单位： | (辽宁师范大学数学学院 大连116029) (北京科技大学计算机与通信工程学院 北京100083) |
| |
| 基金项目： | 国家自然科学基金项目(60875029)资助 |
| |
| 摘 要： | 非负矩阵分解(NMF)是在矩阵非负约束下的一种局部特征提取算法。为了提高识别率,提出了稀疏约束图正则非负矩阵分解方法。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,并将它们整合于单个目标函数中。构造了一个有效的乘积更新算法,并且在理论上证明了该算法的收敛性。在ORL和MIT-CBCL人脸数据库上的实验表明了该算法的有效性。
|
| 关 键 词： | 非负矩阵 图正则化 稀疏编码 |
Graph Regularized Non-negative Matrix Factorization with Sparseness Constraints |
| |
| JIANG Wei,LI Hong,YU Zhen-guo,YANG Bing-ru.Graph Regularized Non-negative Matrix Factorization with Sparseness Constraints[J].Computer Science,2013,40(1):218-220,256. |
| |
| Authors: | JIANG Wei LI Hong YU Zhen-guo YANG Bing-ru |
| |
| Affiliation: | (School of Mathematics, Liaoning Normal University, Dalian 116029, China) (School of Computer & Communication Engineering,University of Science and Technology Beijing,Beijing 100083,China) |
| |
| Abstract: | Nonncgativc matrix factorization(NMI)is based on part feature extraction algorithm which adds nonnegative constraint into matrix factorization. A method called graph regularized non-negative matrix factorization with sparseness constraints(UNMFSC)was proposed for enhancing the classification accuracy. It not only considers the geometric structure in the data representation, but also introduces sparseness constraint to coefficient matrix and integrates them into one single objective function. An efficient multiplicative updating procedure was produced along with its theoretic justificanon of the algorithmic convergence. Experiments on ORI. and MI T-CBCI. face recognition databases demonstrate the effectiveness of the proposed method. |
| |
| Keywords: | Non-negative matrix Graph Regularization Sparse coding |
| 本文献已被 CNKI 等数据库收录！ |
|
点击此处可从《计算机科学》下载全文 |
|
