Abstract
Nonnegative matrix factorization is a linear dimensionality reduction technique used for decomposing high-dimensional nonnegative data matrices for extracting basic and latent features. This technique plays fundamental roles in music analysis, signal processing, sound separation, and spectral data analysis. Given a time-varying objective function or a nonnegative time-dependent data matrix Y(t), the nonnegative factors of Y(t) can be obtained by taking the limit points of the trajectories of the corresponding ordinary differential equations. When the data are time dependent, it is natural to devise factorization techniques that capture the time dependency. To achieve this, one needs to solve continuous-time dynamical systems derived from iterative optimization schemes and construct nonnegative matrix factorization algorithms based on the solution curves. This article presents continuous nonnegative matrix factorization methods based on the solution of systems of ordinary differential equations associated with time-dependent data. In particular, we propose two new continuous-time algorithms based on the Kullback–Leibler divergence and the Amari \(\alpha \)-divergence.
Similar content being viewed by others
References
Lee, D.D., Seung, H.S.: Learning the parts of objects by non-negative matrix factorization. Nature 401(6755), 788–791 (1999)
Lee, D.D., Seung, H.S.: Algorithms for non-negative matrix factorization. In: Proceedings of the advances in neural information processing systems conference, Vol. 13, pp. 556–562. MIT Press, Cambridge (2001)
Chu, M.T., Plemmons, R.J.: Nonnegative matrix factorization and applications. IMAGE Bull. Int. Linear Algebra Soc. 34, 2–7 (2005)
Cichocki, A., Zdunek, R., Phan, A.H., Amari, S.: Nonnegative Matrix and Tensor Factorizations: Applications to Exploratory Multi-Way Data Analysis and Blind Source Separation. Wiley, New York (2009)
Gillis, N.: The why and how of nonnegative matrix factorization. In: Regularization, optimization, kernels, and support vector machines, machine learning and pattern recognition series. Chapman and Hall/CRC, Boca Raton (2014)
Wang, Y.-X., Zhang, Y.-J.: Nonnegative matrix factorization: a comprehensive review. Knowl. Data Eng. IEEE Trans. 25(6), 1336–1353 (2013)
Berry, M.W., Browne, M., Langville, A.N., Pauca, V.P., Plemmons, R.J.: Algorithms and applications for approximate nonnegative matrix factorization. Computat. Stat. Data Anal. 52, 155–173 (2007)
Donoho, D., Stodden, V.: When Does Non-Negative Matrix Factorization Give Correct Decomposition into Parts? MIT Press, Cambridge (2003)
Cichocki, A., Amari, S.: Families of alpha- beta- and gamma- divergences: flexible and robust measures of similarities. Entropy 12(6), 1532–1568 (2010)
Cichocki, A., Cruces, S., Amari, S.: Generalized alpha–beta divergences and their application to robust nonnegative matrix factorization. Entropy 13, 134–170 (2011)
Omlor, L., Slotine, J.J.: Continuous non-negative matrix factorization for time-dependent data. In: Proceedings of the 17th European signal processing conference (2009)
Chu, M.T., Norris, L.K.: Isospectral flows and abstract matrix factorizations. SIAM J. Numer. Anal. 25(6), 1383–1391 (1988)
Chu, M.T., Driessel, K.R.: The projected gradient method for least squares matrix approximations with spectral constraints. SIAM J. Numer. Anal. 27, 1050–1060 (1990)
Chu, M.T.: Numerical linear algebra algorithms as dynamical systems. Acta Numer. 17, 1–86 (2008)
Helmke, U., Moore, J.B.: Optimization and Dynamical Systems. Springer, Berlin (1994)
Chu, M.T., Del Buono, N., Lopez, L., Politi, T.: On the low-rank approximation of data on the unit sphere. SIAM J. Matrix Anal. Appl. 27, 46–60 (2003)
Del Buono, N., Lopez, L.: Geometric integration of manifold of square oblique rotation matrices. SIAM J. Matric Anal. Appl. 23(4), 974–979 (2002)
Del Buono, N., Lopez, L.: Runge–Kutta type methods based on geodesics for systems of ODEs on the Stiefel manifold. BIT Int. Conf. Numer. Math. 41(5), 912–923 (2001)
Del Buono, N., Lopez, L.: Differential approaches for computing Euclidean diagonal norm balanced realizations in control theory. Futur. Gener. Comput. Syst. 19(7), 1155–1163 (2003)
Chu, M.T., Diele, F., Plemmons, R., Ragni, S.: Optimality, computation, and interpretation of nonnegative matrix factorization (2004) (Unpublished)
Luenberger, D.G.: Linear and Nonlinear Programming, 2nd edn. Addison Wesley Publishing Company, Boston (2003)
Lin, C.-J.: Projected gradient methods for non-negative matrix factorization. Neural Comput. 19(10), 2756–2779 (2007)
Trendafilov, N.T.: The dynamical system approach to multivariate data analysis. J. Comput. Gr. Stat. 15(3), 628–650 (2006)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Belachew, M.T., Del Buono, N. A Dynamical System Approach for Continuous Nonnegative Matrix Factorization. Mediterr. J. Math. 14, 14 (2017). https://doi.org/10.1007/s00009-016-0837-y
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00009-016-0837-y