Short summary of the problem:
I am trying to use brms package for the zero-inflated negative family in the Bayesian framework. My aim is to apply the shrinkage method through a horseshoe prior.
To know the importance of variables, I invite experts from different countries and conducted an online survey for recommending whether a predictor candidate is significant (based on the Likert scale). Thus, obtain a weighted score corresponding to each of them.
I want to include these weighted scores as prior information in the model with the horseshoe prior. I tried different ways to solve it but didn’t succeed yet.
The function for defining horseshoe prior in the model:
horseshoe(df = 1, scale_global = 1, df_global = 1, scale_slab = 2, df_slab = 4,
par_ratio = NULL, autoscale = TRUE)
- Operating System: Windows
- brms Version: 2.18.0