Infusion reference page



Coverage of summary likelihood (plain lines) compared to ABC (dashed lines) under the population-size change model of Peter et al. 2010. See Rousset et al. 2017, fig. 2 for details.
Infusion is an R package for Inference using simulation, available on  CRAN

In recent years, simulation methods such as approximate Bayesian computation have extensively been used to infer parameters of population genetic models where the likelihood is intractable. Infusion implements an alternative approach, summary likelihood, that provides a likelihood-based analysis of the information retained in the summary statistics whose distribution is simulated. Rousset et al (2017) show that the method provides confidence intervals with controlled coverage independently of a prior distribution on parameters, in contrast to approximate Bayesian computation. 
gentle introduction with additional examples is also available

This page (C) F. Rousset 2016-present

Main web page