Inferring COVID-19 spreading rates and potential change points for case number forecasts

Jonas Dehning; Johannes Zierenberg; Frank Paul Spitzner; Michael Wibral; Joao Pinheiro Neto; Michael Wilczek; Viola Priesemann

Factor weekends discounted f w 0.6 Inference with the SIR model, as in the main text, but with a weekly modulation of the number of observed cases, to take into account the fewer number of tests performed on weekends. This was modeled by a multiplication of the inferred cases by the absolute value of a sine function (one week period). The offset of the sine function was given a flat prior. The magnitude of the sine function was given a beta-distribution as prior, with mean 0.7 and standard deviation 0.17. A: Time-dependent model estimate of the effective growth rate λ * (t). B: Comparison of daily recorded new cases and the model. Comparison of total recorded cases and the model. C: Model forecast of total cases, based on the inferred λ * (t). D-F: Prior and posterior distributions of all free parameters.