Approximate Bayesian computation for disease outbreaks: Epidemiologists have developed models
for disease spread, but these models often lead to intractable likelihoods. One strategy
for fitting parameters in such models is approximate Bayesian computation (ABC), which
is a simulation-based method for fitting models to data (we will discuss ABC in detail later
in the course).
You will write functions that perform ABC to fit parameters in a model of the spread of a
virus. This will involve:
– Drawing parameters from a prior distribution,
– Drawing data according to a probability model given the parameters,
– Computing a similarity measure between the simulated data and an observed dataset,
– Keeping or discarding the samples based on that similarity.
The goal will be to recreate Figures 3a and 3c in Tony and Stumpf, “Simulation-based model
selection for dynamical systems in systems and population biology”, Bioinformatics (2010)
(available at [login to view URL]), with
the data provided in the supplement to that paper.