Multi-Way, Multi-View Learning

TL;DR

This paper introduces a multi-way, multi-view latent variable model for analyzing high-dimensional, paired data from different domains, extending Bayesian CCA to incorporate multi-way covariates and group-wise metabolite correlations.

Contribution

It extends Bayesian CCA with multi-way covariate integration and a factor analysis for correlated groups, tailored for high-dimensional, paired multi-view data.

Findings

Applied to bioinformatics metabolite data across tissues and conditions.

Demonstrated improved modeling of multi-view, high-dimensional data.

Provided a new framework for multi-way, multi-view analysis in bioinformatics.

Abstract

We extend multi-way, multivariate ANOVA-type analysis to cases where one covariate is the view, with features of each view coming from different, high-dimensional domains. The different views are assumed to be connected by having paired samples; this is a common setup in recent bioinformatics experiments, of which we analyze metabolite profiles in different conditions (disease vs. control and treatment vs. untreated) in different tissues (views). We introduce a multi-way latent variable model for this new task, by extending the generative model of Bayesian canonical correlation analysis (CCA) both to take multi-way covariate information into account as population priors, and by reducing the dimensionality by an integrated factor analysis that assumes the metabolites to come in correlated groups.