Approximate Bayesian Computation

Approximate Bayesian Computation (ABC) is an increasingly popular inference paradigm in applications where traditional (Bayesian or frequentist) inference is difficult because the likelihood function is e.g. computationally costly to evaluate or unavailable in closed, tractab- le form. In the course, we will start by briefly studying traditional Bayesian inference and different kinds of simulation-based inference methods in general. Then we proceed to ABC, covering theoretical and computational aspects including summary statistics for likelihood ap- proximation, posterior sampling methods including rejection, markov chain monte carlo, and population monte carlo methods, as well as model selection. The course topics will to some extent be decided during the course because part of the examination will be through lectures given by the participants. 

The course is given Jan 15, 2018 - Mar 17, 2018 

The lecturer is Magnus Röding / 

Course Data
Type of schedule: 
Schedule to be decided
Credits (ECTS): 
Not scheduled