期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

Nonnegative tensor factorizations using an alternating direction method 被引量：4

Nonnegative tensor factorizations using an alternating direction method
原文传递
导出
摘要 The nonnegative tensor （matrix） factorization finds more and more applications in various disciplines including machine learning, data mining, and blind source separation, etc. In computation, the optimization problem involved is solved by alternatively minimizing one factor while the others are fixed. To solve the subproblem efficiently, we first exploit a variable regularization term which makes the subproblem far from ill-condition. Second, an augmented Lagrangian alternating direction method is employed to solve this convex and well-conditioned regularized subproblem, and two accelerating skills are also implemented. Some preliminary numerical experiments are performed to show the improvements of the new method. The nonnegative tensor （matrix） factorization finds more and more applications in various disciplines including machine learning, data mining, and blind source separation, etc. In computation, the optimization problem involved is solved by alternatively minimizing one factor while the others are fixed. To solve the subproblem efficiently, we first exploit a variable regularization term which makes the subproblem far from ill-condition. Second, an augmented Lagrangian alternating direction method is employed to solve this convex and well-conditioned regularized subproblem, and two accelerating skills are also implemented. Some preliminary numerical experiments are performed to show the improvements of the new method.
作者 Xingju CAI Yannan CHEN Deren HAN
机构地区 School of Mathematical Sciences College of Science
出处 《Frontiers of Mathematics in China》 SCIE CSCD 2013年第1期3-18,共16页 中国高等学校学术文摘·数学（英文）
关键词 Nonnegative matrix factorization nonnegative tensor factorization nonnegative least squares alternating direction method Nonnegative matrix factorization, nonnegative tensor factorization,nonnegative least squares, alternating direction method
分类号 TP391.41 [自动化与计算机技术—计算机应用技术] X523 [环境科学与工程—环境工程]
  • 相关文献

参考文献35

  • 1Acar E, Dunlavy D M, Kolda T G. A scalable optimization approach for fitting canonical tensor decompositions. J Chemometries, 2011, 25(2): 67-86.
  • 2Bader B W, Kolda T G. Algorithm 862: MATLAB tensor classes for fast algorithm prototyping. ACM Trans Math Software, 2006, 32(4): 635 653.
  • 3Bader B W, Kolda T G. Efficient MATLAB computations with sparse and factored tensors. SIAM J Sci Comput, 2007, 30(1): 205 231.
  • 4Benetos E, Kotropoulos C. Non-negative tensor factorization applied to music genre classification. IEEE Trans Audio, Speech, Language Processing, 2010, 18(8): 1955 1967.
  • 5Benthem M H Van, Keenan M R. Fast algorithm for the solution of large-scale non- negativity-constrained least squares problems. J Chemometrics, 2004, 18(10): 441 450.
  • 6Berry M W, Browne M. Email surveillance using non-negative matrix factorization. Comput Math Organization Theory, 2005, 11(3): 249-264.
  • 7Berry M W, Browne M, Langville A N, Pauca V P, Plemmons R J. Algorithms and applications for approximate nonnegative matrix factorization. Comput Statist Data Anal, 2007, 52(1): 155 173.
  • 8Boyd S, Parikh N, Chu E, Peleato B, Eckstein J. Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers. In: Jordan M, ed. Foundations and Trends in Machine Learning, Vol 3. 2011, 1-122 http://www.stanford.edu/-boyd/papers/admm_distr_st ats.html.
  • 9Bro R, Jong S De. 2 fast non-negativity-constrained least squares algorithm. J Chemometrics, 1997, 11(5): 393-401.
  • 10Chen Y, Wang X, Shi C, Lua E K, Fu X M, Deng B X, Li X. Phoenix: a weight-based network coordinate system using matrix factorization. IEEE Trans Network Service Management, 2011, 8(4): 334-347.

同被引文献51

  • 1Zhi-HuaZhou YangYu.Adapt Bagging to Nearest Neighbor Classifiers[J].Journal of Computer Science & Technology,2005,20(1):48-54. 被引量：7
  • 2Tucker L R. Some mathematical notes on three-modefactor analysis[J]. Psychometrika, 1966, 31(3) : 279-311.
  • 3Paatero P,Tapper U. Positive matrix factorization: anon-negative factor model with optimal utilization of er-ror estimates of data values [ J ]. Environmetrics,1994,5(2) : 111-126.
  • 4Lee D D,Seung H S. Learning the parts of objects bynon-negative matrix factorization [ J ]. Nature, 1999,401(6755) : 788-791.
  • 5Hazan T,Polak S, Shashua A. Sparse image codingusing a 3D non-negative tensor factorization[ C ]//10thIEEE International Conference on Computer Vision‘Beijing,2005 : 50-57.
  • 6Morup M, Hansen L K,Amfred S M. Algorithms forsparse nonnegative Tucker decompositions [ J ]. NeuralComputation, 2008,20(8) : 2112—2131.
  • 7Cichocki A, Zdunek R,Phan AH, et al. Alternatingleast squares and related algorithms for NMF and SCAproblems in nonnegative matrix and tensor factorizations[M ]. Chichester,UK: John Wiley & Sons, Ltd,2009: 203 -266.
  • 8Karoui M S,Deville Y,Hosseini S,et al. Blind spatialunmixing of multispectral images : new methods combi-ning sparse component analysis,clustering and non-neg-ativity constraints [ J ]. Pattern Recognition, 2012,45(12): 4263-4278.
  • 9Asaei A, Davies M E, Bourlard H,et al. Computation-al methods for structured sparse component analysis ofconvolutive speech mixtures [ C ] //IEEE InternationalConference on Acoustics, Speech and Signal Process-ing. Kyoto, Japan, 2012: 2425 -2428.
  • 10Kolda T G. Multilinear operators for higher-order de-compositions^ M]. California, USA: Sandia NationalLaboratories, 2006.

引证文献4

二级引证文献18

Frontiers of Mathematics in China
2013年 第1期

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部