A Short Tutorial on Approximate Bayesian Computation and its Applications for Infectious Disease Models

Approximate Bayesian Computation (ABC) originated from the need to address challenging inferential problems in population genetics, where the complexity of the model meant its associated likelihood function could not be evaluated numerically in any practical amount of time. The main idea of ABC is straightforward: generating parameters and the corresponding pseudo-data from the studied model, then keeping only parameters that yield a close match with the observed data. The simplicity of ABC makes the method highly accessible to researchers and has been extensively applied to a wide range of disciplines during the last decade.  

This workshop will provide a basic overview of ABC. Via the statistical software R, we will apply ABC to an infectious disease model to study a COVID-19 data set of the U.S.