``` options(scipen=999) #to avoid scientific notation #setwd("C:/Users/User/OneDrive - ufabc.edu.br/Doutorado/Tentando de novo/Projeto_pos_FAPESP/Leonardo/Cap_2_Tese/") setwd("C:\\Users\\leona\\OneDrive - ufabc.edu.br\\Doutorado\\Tentando de novo\\Projeto_pos_FAPESP\\Leonardo\\Cap_2_Tese") data_neot
Hi Leonardo Matheus Servino it would be best to set your tags to to "Bayesian" or otherwise in the future instead of Packaging or Running then it might be easier to find for people ;). Also the provided reprex is a bit difficult to follow, it would be better to prived an example in Rscript.
The divergent transititions are instabilities in your chains. I am not that familiair with phylo- thing perse and indeed you might adjust your adapt delta or increase the tree-depth, but this is likely not always the "correct" thing to do.
I see you did not specify your priors. There might be some instabilities in your chains because the priors under default are weak priors. If your sample size is small it might be the case that your chains do not mix well. Otherwise, "incorrect" priors, in the sense that your likelihood is not really well aligned with the priors can also result in difficulties within the chains. Meaning that you are somewhere sampling in the tail-area.
I would advice to check your chains via a traceplot and if the current posteriors align with the priors. Increasing the number of itterations is in any cased adviced and thinning the chains could also help. The divergent transitions however are not resolved by this. Yet, without more info I cannot be of any help.
Best,
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation