Hello Everyone,
I wanted to hear thoughts/feedback on the following.
I am fitting a model and 1 out of every 5 fits has divergent transitions(feeding the same data list)
I think it has to do with me not specifying initial values? Would it bias inference if I refit for the divergent fits?
Thoughts, tips would be greatly appreciated.
Hey there!
My opinion on this: If divergent transitions are rare just try to increase adept_delta till they are gone for good. The better option is probably to tweak your model a bit (probably tighter priors will do). It’s preferable to have the model consistently not giving divergent transitions.
Cheers,
Max
Hi Max!
Yes this was on the lines that I was thinking. Thank you for your feedback.