Bayesian Poisson-Randomized Gamma Tensor Factorization with Application to International Trade Flows

DOI10.48550/arXiv.2606.17267MaRDI QIDQ7269949arXivFDO

Authors Aaron Schein, Jie Jian

Publication date 15 June 2026

Published in arXiv

Tldr We propose Bayesian Poisson-randomized Gamma tensor factorization with shared latent structure across modes for modeling sparse semi-continuous multi-way data. A hybrid variational-Monte Carlo algorithm scales posterior inference to large tensors, surfacing dependence in trade flows that pairwise methods miss.

Full work available at URL https://doi.org/10.48550/arXiv.2606.17267

arXiv classification

econ.EMmath.NAstat.APstat.MEstat.ML

This page was built for publication: Bayesian Poisson-Randomized Gamma Tensor Factorization with Application to International Trade Flows

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7269949)