Discrete Exponential-Family Models for Multivariate Binary Outcomes

TL;DR

This paper introduces a flexible, computationally efficient exponential-family model for analyzing multivariate binary outcomes, enabling better hypothesis testing of correlated binary data.

Contribution

The paper presents a novel exponential-family model tailored for multivariate binary data, addressing limitations in scale and interpretability of existing methods.

Findings

Provides a flexible framework for hypothesis testing

Enhances computational efficiency in multivariate binary analysis

Addresses limitations of existing multivariate outcome models

Abstract

Studies that collect multi-outcome data such as tobacco and alcohol use are becoming increasingly common. In principle, multi-outcomes studies investigate the correlations between outcomes, including, causal links and/or joint distributions. Although there are many methods for studying multivariate outcomes, significant limitations regarding scale and interpretation persist. Here we introduce a model based on the exponential-family for discrete binary outcomes that provides a flexible framework for hypothesis testing of multiple binary outcomes in a computationally efficient fashion.