Threshold analysis via weighted posterior predictive distributions (stacking)



I’m working with brms to fit some candidate models that represent different possible patterns of a response variable in respect to a distance measure, and beyond the goal of detecting the most relyable shape, in the case of asymptotic behaviour I’m want to detect at what distance this asymptote is achieved.

I used to do this with a simple randomization test (using the upper values of x as reference), but since I’m going through model averaging via stacking (LOO), want to know if there is a way of use directly the combined weighted predicted distributions to detect this possible threshold.

Any ideas will be very aprecciated.