Computationally Efficient Parameter Estimation for Spatial Individual-Level Models of Infectious Disease Transmission

Date

2020-07

Authors

Ward, Madeline

Journal Title

Journal ISSN

Volume Title

Publisher

University of Guelph

Abstract

Individual-level models (ILMs) incorporate individual-specific covariate information, such as spatial location, to model infectious disease transmission. However, fitting these models with traditional Bayesian methods becomes cumbersome as model complexity or population size increases. To help mitigate this issue, a method for fitting ILMs to aggregate-level data using traditional Metropolis-Hastings Markov chain Monte Carlo and then disaggregating the results to obtain individual-level estimates for epidemic metrics is proposed. This new method is compared to two algorithms within an alternative model-fitting framework, approximate Bayesian computation (ABC). The ABC estimates were generally more accurate but were accompanied by wide credible intervals and in some cases widely varying computation times. The aggregation method was simpler to carry out and had consistent computation times, however its estimates tended to be somewhat biased. The preferred ILM-fitting method would need to be determined case-by-case, based on the priorities of a particular problem.

Description

Keywords

infectious disease modelling, approximate Bayesian computation, individual-level models, data aggregation, epidemic modelling, Markov chain Monte Carlo

Citation