This message was deleted.

Hi @Milim Kim, you can use the

predict_rolled

method. We are additionally working on a

predict_insample

method to make it more simple