Open damonbayer opened 1 month ago
Depending on our faith in the magic land of Julia we may need to think about the stan version which is kept on the log scale for numerical stability: https://github.com/epinowcast/primarycensoreddist/blob/1dbaf5fcacc3b1d83c5752712b1c075cef018eed/inst/stan/primary_censored_dist_analytical_cdf.stan#L27
Aside from the log scale it also expresses the solutions directly as a cdf which is a bit more efficient
See methods in https://github.com/epinowcast/primarycensoreddist/blob/main/R/primary_censored_dist.R