On Rank Selection for Nonnegative Matrix Factorization
Rank selection, i.e. the choice of factorization rank, is the first step in constructing Nonnegative Matrix Factorization (NMF) models. It is a long-standing problem which is not unique to NMF, but arises in most models which attempt to decompose data into its underlying components. Since these models are often used in the unsupervised setting, the rank selection problem is further complicated by the lack of ground truth labels. In this paper, we review and empirically evaluate the most commonly used schemes for NMF rank selection.
Authors
- Ramakrishnan Kannan (ORCID: https://orcid.org/0000-0002-5852-4806)
- Koby Hayashi (ORCID: https://orcid.org/0000-0003-0781-0543)
- Richard Vuduc (ORCID: https://orcid.org/0000-0003-2178-138X)
- Grey Ballard (ORCID: https://orcid.org/0000-0003-1557-8027)
- Srinivas Eswar (ORCID: https://orcid.org/0000-0002-3418-7796)
- Benjamin Cobb (ORCID: https://orcid.org/0000-0002-3214-627X)
- Haesun Park (ORCID: https://orcid.org/0000-0001-6259-7170)
Institutions
- Argonne National Laboratory (US)
- Oak Ridge National Laboratory (US)
- Georgia Institute of Technology (US)
- Pacific Northwest National Laboratory (US)
- Wake Forest University (US)
Publication Details
- Published
- 2024-12-15
- DOI
- https://doi.org/10.1109/bigdata62323.2024.10825324
- Citations
- 1
- Primary Topic
- Face and Expression Recognition
- Type
- article
- Field-Weighted Citation Impact
- 0.25
Funders
- National Science Foundation
- U.S. Department of Energy
- Argonne National Laboratory