S3. Fitting model candidates to low-fidelity D^1j datasets from Anchoring the mean generation time in the SEIR to mitigate biases in ℜ₀ estimates due to uncertainty in the distribution of the epidemiological delays

This appendix illustrates the process of fitting various parameterisations of the SE¹I^jR model (M^1j) to low-fidelity D^1j incidence reports. We mean by parameterisation the decision of categorising parameters as either unknown or assumed. For the unknown parameters, we construct prior distributions, which will be eventually updated in light of the data via HMC sampling, resulting in a posterior distribution. On the other hand, assumed parameters are fixed at their true values.