1

I have a model, the brms code is given below. It is a system of equations (I am estimating demand for two categories of goods). Economic theory tells me that the intercepts have to be restricted to the unit interval. So, I have set my prior, lb, and ub to reflect that. However, I get the error "Error: Argument 'coef' may not be specified when using boundaries." I take this to mean that I have to restrict all my population effects (slope and intercept) to the unit interval or none of them to the unit interval. Is this the case? Is there a way around this?

formula <- mvbf(Foodaway ~ 0 + 
                       Intercept + 
                       log(CPI_FA) + 
                       log(CPI_FH) + 
                       log(CPI_Other) + 
                       I(log(TE)-LogP),
                     Foodhome ~ 0 + 
                       Intercept + 
                       log(CPI_FA) + 
                       log(CPI_FH) + 
                       log(CPI_Other) + 
                       I(log(TE)-LogP),
                     rescor = TRUE)

fit <- brm(formula,
                data = fitdata2017_expanded,
                prior = c(set_prior("beta(1, 1)", class = "b", coef = "Intercept", 
                                resp = "Foodaway", lb = 0, ub = 1),
                          set_prior("beta(1, 1)", class = "b", coef = "Intercept", 
                                resp = "Foodhome", lb = 0, ub = 1),
                          prior("normal(0,5)", class = "b", resp = "Foodaway"),
                          prior("normal(0,5)", class = "b", resp = "Foodhome")), 
                cores = 4,
                chains = 4,
                iter = 5000,
                control = list(max_treedepth = 15))```
Loren P Wagner
  • 11
  • 1
  • I don't know what does that mean, but if you set a Beta prior you don't need to specify the bounds 0 and 1, since a Beta random variable lies between 0 and 1. – Stéphane Laurent May 15 '21 at 15:11
  • Unfortunately, I think that is not quite true. If you set the prior without the lb and ub arguments you get this warning:1: It appears as if you have specified a lower bounded prior on a parameter that has no natural lower bound. If this is really what you want, please specify argument 'lb' of 'set_prior' appropriately. Warning occurred for prior b_Foodaway_Intercept ~ beta(1,1) b_Foodhome_Intercept ~ beta(1,1) You get the same thing for the upper bound. – Loren P Wagner Oct 17 '21 at 00:17

0 Answers0