Any implementation of "Bayesian" dominance analysis (like relative importance) so far?

General
,
1

Hi!

For my research, I frequently run GLMM (e.g., ordinal (cumulative) regression with simple random effect structure [random intercepts of a variable]), and would like to know if there is any implementation of 'Dominance analysis" or relative importance (like implemented in the domir package or relimpo package). I think that my question is general enough not to involve specific codes, but if more information is necessary to assess the context, please do let me know.

Thank you very much!

2

This might not be exactly what your looking for but you could have a look at what @avehtari and his research group has done with ProjPred. The goal of their research is variable selection but it is also based on the contribution of different variables.