Supplementary methods
Data
The public repository CSSEGISandData/COVID-19 contains the list of confirmed, recovered and death cases updated every day. To get the number of infected individual at each day we consider the difference between confirmed and recovered or deats cases.
Likelihood calculation
The data likelihood is calculated as
where $X_{it}(h,x_0,k,g)$ is the average number of infected cases in country $i$ at time $t$, which was obtained by numerical integration of the differential equations of the model.
Prior distributions
As prior distributions we used gamma distributions as follows
Sampling strategy
To draw samples from the posterior distribution of the model parameters we employed a standard Metropolis-Hastings scheme with a proposal obtained by combining a gamma distribution with equal scale and shape (mean at 1 to allow for local sampling around the current value) and the corresponding prior to allow for larger transitions: