Dear all, I am new in Bayesian statistics and I have a very simple, basic question. I’m modeling count data and I’m trying to use Bayesian statistics through brms package. My response variable has underdispersion and so “family=negbinomial()” is not adequate. In the frequentist approach, I always u…

You mean the COM-Poisson distribution? It is available, but not openly exposed ;-) You can access it via family = brmsfamily("com_poisson")

Brms and Conway-Maxwell-Poisson distribution

andre.pfeuffer January 22, 2019, 11:04am 6

The gamma-count distribution handles under-dispersion as well and can be constructed more easily. I did one time in Stan (not BRMS).

1 Like

Topic		Replies	Views
Post-processing com_poisson model in brms brms	5	638	January 1, 2020
Mixture model in brms with known mixture locations brms	2	930	July 16, 2019
Divergence with COM_poisson model brms rstan , fitting-issues , brms	2	719	March 9, 2022
Poisson or Neg. Binomial with non-integer response Modeling rstan , specification , brms	2	1616	November 29, 2022
Brms model with com_poisson family: interaction involving a categorical variable brms	4	1164	April 24, 2019