Regularized Alternating Least Squares Algorithms for Non-negative Matrix/Tensor Factorization

Abstract

Nonnegative Matrix and Tensor Factorization (NMF/NTF) and Sparse Component Analysis (SCA) have already found many potential applications, especially in multi-way Blind Source Separation (BSS), multi-dimensional data analysis, model reduction and sparse signal/image representations. In this paper we propose a family of the modified Regularized Alternating Least Squares (RALS) algorithms for NMF/NTF. By incorporating regularization and penalty terms into the weighted Frobenius norm we are able to achieve sparse and/or smooth representations of the desired solution, and to alleviate the problem of getting stuck in local minima. We implemented the RALS algorithms in our NMFLAB/NTFLAB Matlab Toolboxes, and compared them with standard NMF algorithms. The proposed algorithms are characterized by improved efficiency and convergence properties, especially for largescale problems.

DOI: 10.1007/978-3-540-72395-0_97

Extracted Key Phrases

4 Figures and Tables

05101520072008200920102011201220132014201520162017
Citations per Year

74 Citations

Semantic Scholar estimates that this publication has 74 citations based on the available data.

See our FAQ for additional information.

Cite this paper

@inproceedings{Cichocki2007RegularizedAL, title={Regularized Alternating Least Squares Algorithms for Non-negative Matrix/Tensor Factorization}, author={Andrzej Cichocki and Rafal Zdunek}, booktitle={ISNN}, year={2007} }