Hello. Is it possible to do tobit type II in brms?
I can do tobit type I with the following R codes using and brms.
library(dplyr)
library(brms)
library(cmdstanr)
# simulate the initial data
n <- 4000
set.seed(100)
x=rnorm(n, mean=1,sd=2)
sim_d <- tibble(y = rnorm(n, mean = .5+.5*x, sd = 1))
# create the censoring
sim_d <- sim_d %>%
mutate(y1 = if_else(y < 0, 0, y),
cen = if_else(y < 0, "left", "none"))
sim_d$x<-x
fit <-
brm(
y1 | cens(cen) ~ x,
prior = c(prior(normal(1, 3), class = "b"),
prior(normal(0, 3), class = sigma)),
data = sim_d,
family = gaussian,
seed = 101,
chains=4, cores=4, backend="cmdstanr")
summary(fit)
Appreciate any suggestions.