I wrote an article on how one might go about combining multiple machine learning models in Ruby. What’s become apparent is that it seems like one may eventually need to leverage multiple languages:
https://lwflouisa.github.io/blog/Multiple-Algorithms-For-Multiple-Reference-Frames
d_baysian is a hybrid ML structure that uses the input from a statistical model, to determine the result for a decision tree, that then uses that decision tree result to come up with a solution in an if then else. But it will still be a separate method under a frame of reference class.