Specifying priors in mediation model

Modeling
1

Hi,

I am wondering if anyone knows how I can specify priors by model/formula in something like this brms mediation model:

y_model ← bf(log_a_markdown ~ 1 + lead_time_std + us_penetration + china_penetration + indonesia_penetration + canada_penetration)
m_model ← bf(lead_time_std ~ 1 + us_penetration + china_penetration + indonesia_penetration + canada_penetration)

a_md_mod ← brm(data = df_home,
family = gaussian,
y_model + m_model + set_rescor(F),
seed = 1555211, iter = 5000, warmup = 500, cores = 4, chains = 4) # we do not want residual correlation

2

Hi

you can use the get_prior function to see what brms is asking from each parameter

prior <- get_prior(y_model + m_model + set_rescor(F), 
data = df_home,
family = gaussian)

From there you can modify the prior object to change them for each part of model