Abstract
Standard nonnegative matrix factorization extracts nonnegative bases for nonnegative representation, which, however, considers only features, not commonness. In addition, the standard NMF is an unsupervised learning method that cannot fully utilize label information if it exists. In this paper, we present a semi-supervised commonness NMF technique that incorporates samples’ commonness and label information into the optimization model. Naturally, the commonness vector should be constrained by nonnegativity and will degenerate to zero if no commonness exists. We develop a multiplicative update rule to solve the model, which has properties comparable to those of the standard NMF with automatic satisfaction of the nonnegativity constraints, monotonicity without the need for any adjustable learning rate and a low computational overhead. Through experiments on the standard databases, we analyze the behavior of the proposed method, which exhibits a performance that is favorably superior with respect to commonness extraction and clustering accuracy.
Similar content being viewed by others
Notes
Lee et al. introduced two weighted matrices to handle missing entries and labels of the data. The former may be used to all NMF methods, which, however, is not considered in the scope of this paper. And the latter is incorporated to the matrix L for convenience.
References
Lee DD, Seung HS (1999) Learning the parts of objects by non-negative matrix factorization. Nature 401:788–791
Lee DD, Seung HS (2000) Algorithms for non-negative matrix factorization. In: Advances in neural information processing systems, pp. 556–562
Hoyer P (2004) Non-negative matrix factorization with sparseness constraints. J Mach Learn Res 5:1457–1469
Dhillon IS, Sra S (2005) Generalized nonnegative matrix approximations with bregman divergences. In: Proceedings of the neural information processing systems, pp. 283–290
Cichocki A, Zdunek R, Amari S (2006) Csiszar’s divergences for non-negative matrix factorization: family of new algorithms. Lect Notes Comput Sci 3889:32–39
Cichockia A, Lee H, Kimb YD, Choi SJ (2008) Non-negative matrix factorization with \(\alpha \)-divergence. Pattern Recogn Lett 29:1433–1440
Fevotte C, Bertin N, Durrieu J (2009) Nonnegative matrix factorization with the itakura-saito divergence: with application to music analysis. Neural Computat 2:793–830
Cho Y, Saul LK (2011) Nonnegative matrix factorization for semi-supervised dimensionality reduction. arXiv:1112.3714
Mohammadiha N, Smaragdis P, Leijon A (2007) Supervised and unsupervised speech enhancement using nonnegative matrix factorization. IEEE Trans Audio Speech 21:2140–2151
Wang D, Gao X, Wang X (2016) Semi-supervised nonnegative matrix factorization via constraint propagation. IEEE Trans Cyber 46:233–244
Lee H, Yoo J, Choi S (2010) Semi-supervised nonnegative matrix factorization. IEEE Signal Process Lett 17:4–7
Cichocki A, Amari S, Zdunek R, Kompass P, Hori G, He Z (2006) Extended smart algorithms for non-negative matrix factorization. Lect Notes Artif Intell 4029:548–562
Zdunek R, Cichocki A (2007) Gibbs regularized nonnegative matrix factorization for blind separation of locally smooth signals. In: Proceedings of the 15th IEEE international workshop on nonlinear dynamics of electronic systems (NDES07), pp. 317–320
Cai D, He X, Han J, Huang T (2011) Graph regularized nonnegative matrix factorization for data representation. IEEE Trans Pattern Anal Mach Intell 33:1548–1560
Liu H, Wu Z, Li X, Cai D, Huang TS (2012) Constrained nonnegative matrix factorization for image representation. IEEE Trans Pattern Anal Mach Intell 34:233–244
Wang C, Yan S, Zhang L, Zhang H (2009) Non-negative semi-supervised learning. In: Proceedings of the twelfth international conference on artificial intelligence and statistics, pp 575–582
Ma H, Jia M, Shi YK, Hao Z, Liu Y, Jin R, Yang L (2014) Semi-supervised multi-label learning by constrained non-negative matrix factorization. In: proceedings of the national conference on artificial intelligence 18th innovative applications of artificial intelligence conference, Boston, pp. 511–516
Kannan R, Ballard G, Park H (2015) A high-performance parallel algorithm for nonnegative matrix factorization. arXiv:1509.09313
Gillis N, Kumar A (2015) Exact and heuristic algorithms for semi-nonnegative matrix factorization. SIAM J Matrix Anal Appl 36:1404–1424
Gillis N, Vavasis SA (2015) Semidefinite programming based preconditioning for more robust near-separable nonnegative matrix factorization. SIAM J Opt 25:677–698
Vandaele A, Gillis N, Lei Q, Zhong K, Dhillon I (2016) Efficient and non-convex coordinate descent for symmetric nonnegative matrix factorization. IEEE Trans Signal Process 64:5571–5584
Bhattacharya C, Goyal N, Kannan R, Pani J (2016) Non-negative matrix factorization under heavy noise. In: Proceedings of the 33rd international conference on machine learning, New York, pp. 1426–1434
Samaria F, Harter A (1994) Parameterisation of a stochastic model for human face identification. In: Proceedings of 2nd IEEE workshop applications of computer vision, pp. 5–7
Bellhumer PN, Hespanha J, Kriegman D (1997) Eigenfaces vs. fisherfaces: recognition using class specific linear projection. IEEE Trans Pattern Anal Mach Intell 17:711–720
Nene SA, Nayar SK, Murase H (1996) Columbia object image library (COIL-100), Technical reports, CUCS-006-96, February, 1996
Lichman M (2013) UCI machine learning repository. University of California, School of Information and Computer Science, Irvine. http://archive.ics.uci.edu/ml
Odewahn S, Stockwell E, Pennington R, Humphreys R, Zunach W (1992) Automated star/galaxy discrimination with neural networks. Astron J 103:318–331
Lovasz L, Plummer M (1986) Matching theory, Amsterdam. Elsevier, The Netherlands
Stone M (1974) Cross-validatory choice and assessment of statistical predictions. J R Stat Soc 36:111–147
Acknowledgements
This work was supported by the National Natural Science Foundations of China (61302013, 61372014, 61374015), Science and Technology Plan of Liaoning Province of China (2014305001) and Fundamental Research Funds for the Central Universities of China (N141008001).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Teng, Y., Qi, S., Dai, Y. et al. Semi-supervised Nonnegative Matrix Factorization with Commonness Extraction. Neural Process Lett 45, 1063–1076 (2017). https://doi.org/10.1007/s11063-016-9565-3
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11063-016-9565-3