计算机与数字工程2017，Vol.45Issue(6)：1019-1024,6.DOI:10.3969/j.issn.1672-9722.2017.06.002
基于非负矩阵分解的低秩矩阵恢复模型
Low-rank Matrix Recovery Model Based on Non-negative Matrix Factorization
摘要
Abstract
To overcome the shortage of large-scale nuclear matrix singular value decomposition existing in low-rank matrix re?covery model,the paper proposed low-rank matrix recovery model based on non-negative matrix factorization. Non-negative matrix factorization(NMF) applied to the low-rank matrix,which could quickly deal with the problem of the decomposition matrix of low-rank and avoid large-scale nuclear matrix singular value decomposition. Then the algorithm used alternarting directions method of multipliers(ADMM). ADMM divided the global problem into partial sub-problems. Each sub-problem used Lagrange multipliers to solve low rank matrix and sparse matrix. Experimental results in ORL,AL_Gore and Windows databases showed that low-rank re?covery model based NMF has higher recognition rate,better reduction rank and lower the complexity of the algorithm than other tra?ditional low-rank recovery model.关键词
非负矩阵分解/低秩矩阵恢复/多乘子交替迭代法/奇异值分解/图像识别Key words
non-negative matrix factorization(NMF)/low-rank matrix recovery/alternating directions method of multipli⁃ers/singular value decomposition(SVD)/image recognition分类
信息技术与安全科学引用本文复制引用
徐梦珂,许道云,魏明俊..基于非负矩阵分解的低秩矩阵恢复模型[J].计算机与数字工程,2017,45(6):1019-1024,6.基金项目
国家自然科学基金项目(编号:61262006,61540050) （编号:61262006,61540050）
贵州省重大应用基础研究项目(编号:黔科合JZ字[2014]2001) （编号:黔科合JZ字[2014]2001）
贵州省科技厅联合基金(编号:黔科合LH字[2014]7636)资助. （编号:黔科合LH字[2014]7636）
