I'm looking for a "voting” classifier which is obtained by splitting a training set into k fragments, building a C4.5-tree on each fragment separately, and using a majority vote of the trees as the output.
What is the the model I have to choose? (Bagging CART/ Boosted C5.0/ Random Forest or Gradient Boosted Machine) Is the following model fits my need where the number of trials equals k:
fit <- C5.0(Species~., data=iris, trials=10)
Or do I have to do any changes for meeting my need of voting and splitting the DB into k fragments? How can I implement it?
