Title: Approximate Bayesian Computation and history matching for inference in infectious disease systems
Authors: Trevelyan McKinley - University of Exeter (United Kingdom) [presenting]
Abstract: Complex mathematical models are being increasingly used to inform decision making. Adequately capturing key sources of uncertainty is important to produce robust predictions and reduce the probability of making poor decisions. Approximate Bayesian Computation (ABC) and other simulation-based inference methods are becoming increasingly popular for inference in complex systems, particularly ones where the likelihood function is intractable. This is due to their relative ease-of-implementation compared to alternative approaches, since they require only the means to simulate from the underlying complex model. However, despite their utility, scaling simulation-based methods to fit large-scale systems introduces a series of additional challenges that hamper robust inference. Here we use a real-world model of HIV transmission - that has been used to explore the potential impacts of potential control policies in Uganda - to illustrate some of these key challenges when applying ABC methods to high dimensional, computationally intensive models. We then discuss an alternative approach - history matching with emulation - that aims to address some of these issues and conclude with a comparison between these different methodologies.