Affordable Access

deepdyve-link
Publisher Website

Negative Binomial Matrix Factorization

Authors
  • Gouvert, Olivier
  • Oberlin, Thomas
  • Févotte, Cédric
Publication Date
Jan 01, 2020
Identifiers
DOI: 10.1109/LSP.2020.2991613
OAI: oai:HAL:hal-02871905v1
Source
HAL-Descartes
Keywords
Language
English
License
Unknown
External links

Abstract

We introduce negative binomial matrix factoriza-tion (NBMF), a matrix factorization technique specially designed for analyzing over-dispersed count data. It can be viewed as an extension of Poisson factorization (PF) perturbed by a multiplicative term which models exposure. This term brings a degree of freedom for controlling the dispersion, making NBMF more robust to outliers. We describe a majorization-minimization (MM) algorithm for a maximum likelihood estimation of the parameters. We provide results on a recommendation task and demonstrate the ability of NBMF to efficiently exploit raw data.

Report this publication

Statistics

Seen <100 times