BayesFlow Benchmarks

SIR

Dynamical systems represent paradigmatic use cases for SBI. SIR is an influential epidemiological model describing the dynamics of the number of individuals in three possible states: susceptible S, infectious I, and recovered or deceased, R. We infer the contact rate $\beta$ and the mean recovery rate $\gamma$, given observed infection counts I at 10 evenly-spaced time points.

Description from Lueckmann et. al, 2021: Benchmarking Simulation-Based Inference

Posterior Samples

Approximation

Reference

Details