Random Forest (page 2 of 5)

The term random forest was first coined by Tin Kam Ho of Bell Labs in 1995. The method combines Breiman's "bagging" idea and the random selection of features. Each tree in the forest is typically made with a different random bootstrapped sample of the training data and a random set of features. A fully "random" random forest builds each tree with a random set of features and a random bootstrapped sample of the training data. The observations not used forms an “out of bag” set for each tree in the forest. These “out of bag” observations are used for validation in a similar manner to cross-validation to validate each tree.