Grinsztajn et al. (2020) use half normal priors on beta and gamma. In particular, mean_beta = 2, sd_beta = 1 and mean_gamma=0.4 and sd_gamma=0.5.
Rupp et al. (2021) use uniform priors on the (log) rates (!!!!). Here's a quote:
We do a full Bayesian analysis for parameter pairs (log α, log β) with a uniform prior. Following Ho, Crawford, et
al. (2018) we sample from the joint posterior using a Hamiltonian Monte Carlo (HMC) scheme (Duane et al. 1987; Neal
2011).
Volz & Silveroni, (2018):
R_0 = beta/gamma, prior:log-normal(0, sd = 1).Lam & Suchard (2018):
R_0 = N*beta/gamma, priors: gamma and beta have alog-normal(0, 100)prior.N = 261.McKinley et al (2019) Exp(1), i.e. Gamma(1, 1) on all rates.
Grinsztajn et al. (2020) use half normal priors on
betaandgamma. In particular,mean_beta = 2,sd_beta = 1andmean_gamma=0.4andsd_gamma=0.5.Rupp et al. (2021) use uniform priors on the (log) rates (!!!!). Here's a quote: