Random Forest

Random Forests have been used as effective ensemble models for classification. We present in this paper a new type of Random Forests (RFs) called Red(uced)-RF that adopts a new dynamic data reduction principle and a new voting mechanism called Priority Vote Weighting (PV) which improve accuracy, execution time and AUC values compared to Breiman’s RF. Red-RF also shows that the strength of a random forest can increase without noticeably increasing correlation between the trees. We then compare performance of Red-RF and Breiman’s RF in 8 experiments that involve classification problems with datasets of different sizes. Finally, we conduct two additional experiments that involve considerably larger datasets with one million points in each.