Hello, I’m using anpan to analysis the correlation between phylogenic and host phenotype. I tried to use multi categorical variables as my outcome but that seems doesn’t work. Is there any way to solve this problem? ANY valuable answer is accepted
Sorry, anpan doesn’t and won’t support categorical outcome variables. Only binary (Bernoulli) and numeric (Gaussian) outcomes are supported.
If you’re sufficiently motivated, adapting the model for a categorical outcome would be doable if you familiarize yourself Stan and the multi-logit regression section of the user guide. The leave-one-out model comparison calculations would be a big jump in implementation difficulty though. That would require some tricky multivariate numerical integrals.
THANKs for your valuable answer!!