Format

Send to

Choose Destination

See 1 citation:

Neural Comput. 2016 Aug;28(8):1663-93. doi: 10.1162/NECO_a_00853. Epub 2016 Jun 27.

A Quasi-Likelihood Approach to Nonnegative Matrix Factorization.

Author information

1
Department of Biostatistics and Bioinformatics, Fox Chase Cancer Center, Temple University Health System, Philadelphia, PA 19111, U.S.A. karthik.devarajan@fccc.edu.
2
School of Biomedical Sciences, Chinese University of Hong Kong, Shatin, NT, Hong Kong vckc@cuhk.edu.hk.

Abstract

A unified approach to nonnegative matrix factorization based on the theory of generalized linear models is proposed. This approach embeds a variety of statistical models, including the exponential family, within a single theoretical framework and provides a unified view of such factorizations from the perspective of quasi-likelihood. Using this framework, a family of algorithms for handling signal-dependent noise is developed and its convergence proved using the expectation-maximization algorithm. In addition, a measure to evaluate the goodness of fit of the resulting factorization is described. The proposed methods allow modeling of nonlinear effects using appropriate link functions and are illustrated using an application in biomedical signal processing.

PMID:
27348511
PMCID:
PMC5549860
DOI:
10.1162/NECO_a_00853
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Atypon Icon for PubMed Central
Loading ...
Support Center