Bayesian Modeling and Prediction of Generalized Contact Matrices

TL;DR

This paper introduces a Bayesian framework for inferring generalized contact matrices in infectious disease modeling, accommodating high-dimensional data and structural constraints, with practical applications demonstrated on real datasets.

Contribution

It presents a novel Bayesian modeling approach that extends contact matrices beyond age, leveraging tensor structures and smoothing for stable, high-dimensional estimation.

Findings

The framework effectively handles missing feature data in contact matrices.

Simulation studies show improved estimation accuracy over existing methods.

Application to US and German datasets demonstrates practical utility.

Abstract

Social contact matrices are essential tools in infectious disease epidemiology as they quantify close-range human contact patterns which directly drive the transmission of airborne infectious diseases. In this work we propose a Bayesian modeling framework for inferring generalized contact matrices which stratify contact matrices beyond contemporary age dimensions. The model is designed to satisfy fundamental structural assumptions of contacts while leveraging tensor structures and smoothing constraints to make high-dimensional matrix estimation computationally feasible and statistically stable. We discover a link between multi-dimensional matrix stratification subject to structural constraints with the theory of contingency tables. This enables us to approach a challenging missing-data problem commonly encountered in real-world analysis where feature information on the contacts is unobserved. We benchmark the framework against existing methods through simulation studies and illustrate the framework's practical utility through two real-world datasets: BICS (United States) and COVIMOD (Germany). Our models are implemented in an open-source Python package to facilitate adoption in the wider scientific community.