A New Family of Poisson Non-negative Matrix Factorization Methods Using the Shifted Log Link
Eric Weine [email protected], Peter Carbonetto [email protected], Rafael A. Irizarry [email protected], Matthew Stephens [email protected] · 2026
Read A New Family of Poisson Non-negative Matrix Factorization Methods Using the Shifted Log Link by Eric Weine [email protected], Peter Carbonetto [email protected], Rafael A. Irizarry [email protected], Matthew Stephens [email protected] (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-05845.txt.
Description
Eric Weine and 3 co-authors introduce in 2026 a new family of Poisson non-negative matrix factorization methods that let data parts combine multiplicatively as well as additively. The method decomposes count data into interpretable parts. Standard versions assume parts simply add, which fits some datasets but not others.
A shifted-log link function, controlled by one tuning parameter, interpolates between additive and multiplicative combination. The authors provide a maximum-likelihood fitting algorithm and a faster approximation whose cost scales with the number of nonzero entries, suiting large sparse data. Examples on varied real datasets show the link choice can change results substantially. In some settings the shifted-log version produced more interpretable decompositions than the standard additive model.
Catalog record
| CREATOR | Eric Weine [email protected], Peter Carbonetto [email protected], Rafael A. Irizarry [email protected], Matthew Stephens [email protected] |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-05845