Algorithms for Nonnegative Matrix Factorization with the Kullback–Leibler Divergence

被引:0
作者
Le Thi Khanh Hien
Nicolas Gillis
机构
[1] Université de Mons,Department of Mathematics and Operational Research, Faculté Polytechnique
来源
Journal of Scientific Computing | 2021年 / 87卷
关键词
Nonnegative matrix factorization; Kullback–Leibler divergence; Poisson distribution; Algorithms;
D O I
暂无
中图分类号
学科分类号
摘要
Nonnegative matrix factorization (NMF) is a standard linear dimensionality reduction technique for nonnegative data sets. In order to measure the discrepancy between the input data and the low-rank approximation, the Kullback–Leibler (KL) divergence is one of the most widely used objective function for NMF. It corresponds to the maximum likehood estimator when the underlying statistics of the observed data sample follows a Poisson distribution, and KL NMF is particularly meaningful for count data sets, such as documents. In this paper, we first collect important properties of the KL objective function that are essential to study the convergence of KL NMF algorithms. Second, together with reviewing existing algorithms for solving KL NMF, we propose three new algorithms that guarantee the non-increasingness of the objective function. We also provide a global convergence guarantee for one of our proposed algorithms. Finally, we conduct extensive numerical experiments to provide a comprehensive picture of the performances of the KL NMF algorithms.
引用
收藏
相关论文
共 76 条
[1]  
Ang AMS(2019)Accelerating nonnegative matrix factorization algorithms using extrapolation Neural Comput. 31 417-439
[2]  
Gillis N(2010)Proximal alternating minimization and projection methods for nonconvex problems: an approach based on the Kurdyka-Łojasiewicz inequality Math. Oper. Res. 35 438-457
[3]  
Attouch H(2017)A descent lemma beyond Lipschitz gradient continuity: first-order methods revisited and applications Math. Oper. Res. 42 330-348
[4]  
Bolte J(2014)Proximal alternating linearized minimization for nonconvex and nonsmooth problems Math. Program. 146 459-494
[5]  
Redont P(2011)A first-order primal-dual algorithm for convex problems with applications to imaging J. Math. Imaging Vis. 40 120-145
[6]  
Soubeyran A(2012)On tensors, sparsity, and nonnegative factorizations SIAM J. Matrix Anal. Appl. 33 1272-1299
[7]  
Bauschke HH(2017)Visualizing the structure of RNA-SEQ expression data using grade of membership models PLoS Genet. 13 e1006599-3927
[8]  
Bolte J(2008)On the equivalence between non-negative matrix factorization and probabilistic latent semantic indexing Comput. Stat. Data Anal. 52 3913-213
[9]  
Teboulle M(2002)Benchmarking optimization software with performance profiles Math. Program. 91 201-830
[10]  
Bolte J(2009)Nonnegative matrix factorization with the Itakura-Saito divergence: with application to music analysis Neural Comput. 21 793-2456