Continuous binomial, negative binomial and Poisson

[Max_Mantei]

Maybe as an alternative to this:

Tweedie Likelihood (compound Poisson-gamma) in Stan

I’ve been trying to fit a Tweedie likelihood model in Stan that allows for inflated zeroes. In case it is useful to anyone, I have reproduced what I have done here. I think the efficiency of the code could probably be improved, but otherwise it appears to work. I have implemented the R “tweedie” package functions (Dunn 2017) which work, but are quite slow. The strategy is to split the data into 3 groups: i) y==0, ii) low y (dependent variable) values for which the series sum can be used and iii…

?