Excelent alternative to use churdle in bayesmh. Is there a way to implement a glm model with family(gamma) and link(identity) using bayesmh?? I assume one would have to write the log-likelihood function as well, but there lies the challenge. Have anyone written that function somewhere?
↧