How to convert hurdle lognormal model posterior values to actual values of y variable?

Multiple intercepts in posterior cause confusion for me.

Data
period - factor, 3 levels, I, II, III
region - factor, 2 levels, A, B

Model
`m = brm(bf(hours ~ period + (period | region), hu ~ period + (period | region)), data = data, family = hurdle_lognormal())`

Could you please tell me how to convert posterior values from hurdle_lognormal model to

1. received hours (lognormal part) for region A and period I
2. probability of zero therapy (binomial part) for region B and period III

Model posterior
` post = posterior_samples(m)`

Column names of posterior
“b_Intercept”
“b_hu_Intercept”
“b_periodII”
“b_periodIII”
“b_hu_periodII”
“b_hu_periodIII”
“sd_county__Intercept”
“sd_county__periodII”
“sd_county__periodIII”
“sd_county__hu_Intercept”
“sd_county__hu_periodII”
“sd_county__hu_periodIII”
“cor_county__Intercept__periodII”
“cor_county__Intercept__periodIII”
“cor_county__periodII__periodIII”
“cor_county__hu_Intercept__hu_periodII”
“cor_county__hu_Intercept__hu_periodIII”
“cor_county__hu_periodII__hu_periodIII”
“sigma”
“r_county[regionB,Intercept]”
“r_county[regionA,Intercept]”
“r_county[regionB,periodII]”
“r_county[regionA,periodII]”
“r_county[regionB,periodIII]”
“r_county[regionA,periodIII]”
“r_county__hu[regionB,Intercept]”
“r_county__hu[regionA,Intercept]”
“r_county__hu[regionB,periodII]”
“r_county__hu[regionA,periodII]”
“r_county__hu[regionB,periodIII]”
“r_county__hu[regionA,periodIII]”
“lp__”

My own solution for question 1: region A, period I
lognormal model intercept + regionA intercept?
`exp((b_Intercept + r_county[regionA,Intercept) + 1/2 * (sigma)**2))`

My own solution for question 2: region B, period III
hurdle model intercept + regionB period III?
`inv_logit_scaled(b_hu_Intercept + r_county__hu[regionB,periodIII)`

Hey there!

Not a `brms` expert here, but…

I think you can just use `posterior_predict` with the `newdata` argument set to a df with the specific cases, right?

Cheers,
Max