using trained weights of learners in a boosted classification ensemble to predict probability

Hello all,

I am wondering how to use the trained weights of each weak learner with the predict function for boosted classification ensemble. I think the predict function by default takes the averaged weighted sum of the learners, rather than than giving weight to some rather than others based on their trained weights. How can I calculate predictions giving more weight to learners with higher trained weight scores?

I would really appreciate any advice on this please!