rstanarm - updating priors with binary data without any events

120 views Asked by At

rstanarm - updating priors with binary data without any events

Hi, I started exploring the rstanarm package while taking a first course in probability. My question is not for an assignment, but more out of curiosity after I read https://cran.r-project.org/web/packages/rstanarm/vignettes/binomial.html

Imagine that in the example provided in this vignette (see code below) the switch variable had only the first 10 observations and no events (i.e., all values are zero). Attempting to run this model without any events generates an error, but as far as I understand there should be a way to update a prior even if there are no instances.

My question: Is it possible to use rstanarm to model a case with 10 observations with no events (all rows for switch with a zero value)?

 t_prior <- student_t(df = 7, location = 0, scale = 2.5)
 fit1 <- stan_glm(switch ~ dist100, data = wells, 
             family = binomial(link = "logit"), 
             prior = t_prior, prior_intercept = t_prior,  
             chains = 4, cores = 2, seed = 1234, iter = 100)
0

There are 0 answers