I have created a mixed effect bayesian GLM using rstanarm. I have a few parameters that I suspect to have correlation (or possibly collinearity) issues from looking at a simple correlation matrix. I then used the function check_collinearity on my model to assess the VIFs and found VIF values to be high (over 10 for some parameters). However, I was wondering if VIFs are still valid metrics in a bayesian framework, or if there are other preferable forms of describing collinearity? If VIFs are not used in Bayesian frameworks, why?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.