Tensor decompositions and sparse log-linear models

DOI10.1214/15-AOS1414zbMATH Open1367.62180arXiv1404.0396OpenAlexW2963574674WikidataQ50050916 ScholiaQ50050916MaRDI QIDQ524449FDOQ524449

Authors: James E. Johndrow, Anirban Bhattacharya, David Dunson 

Publication date: 2 May 2017

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Contingency table analysis routinely relies on log linear models, with latent structure analysis providing a common alternative. Latent structure models lead to a low rank tensor factorization of the probability mass function for multivariate categorical data, while log linear models achieve dimensionality reduction through sparsity. Little is known about the relationship between these notions of dimensionality reduction in the two paradigms. We derive several results relating the support of a log-linear model to the nonnegative rank of the associated probability tensor. Motivated by these findings, we propose a new collapsed Tucker class of tensor decompositions, which bridge existing PARAFAC and Tucker decompositions, providing a more flexible framework for parsimoniously characterizing multivariate categorical data. Taking a Bayesian approach to inference, we illustrate advantages of the new decompositions in simulations and an application to functional disability data.

Full work available at URL: https://arxiv.org/abs/1404.0396

Recommendations

• Bayesian factorizations of big sparse tensors
• Provable Sparse Tensor Decomposition
• Tensor decompositions for learning latent variable models
• Tensor Decompositions for Learning Latent Variable Models (A Survey for ALT)
• On Tensors, Sparsity, and Nonnegative Factorizations

zbMATH Keywords

Bayesiancontingency tablelatent class analysissparsitygraphical modelhigh-dimensionalcategorical datalow rankPARAFACTucker

Mathematics Subject Classification ID

Bayesian inference (62F15) Characterization and structure theory for multivariate probability distributions; copulas (62H05) Contingency tables (62H17)

Cited In (13)

• Exploring dependence between categorical variables: benefits and limitations of using variable selection within Bayesian clustering in relation to log-linear modelling with interaction terms
• Detecting structural changes in longitudinal network data
• Bayesian inference on group differences in multivariate categorical data
• Title not available (Why is that?)
• An optimal statistical and computational framework for generalized tensor estimation
• Modeling association in microbial communities with clique loglinear models
• Integrating sample similarities into latent class analysis: a tree‐structured shrinkage approach
• Title not available (Why is that?)
• Sparse random tensors: concentration, regularization and applications
• Tensor Decompositions for Learning Latent Variable Models (A Survey for ALT)
• Composite mixture of log-linear models with application to psychiatric studies
• Bayesian nonparametric multiway regression for clustered binomial data
• Cross: efficient low-rank tensor completion

This page was built for publication: Tensor decompositions and sparse log-linear models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q524449)