Current implementation of majority voting prediction for decision trees implicitly assumes that all trees are equally weighted. This should be updated to use the particle weights, probably by taking in the MCMCOutput/SMCOutput instead of the list of trees and weight the trees' votes accordingly