Random Forest
A random forest is a non-parametric ensemble model using decision trees.
Contents
Description
A random forest is a modification of bagging methods.
A sample of features are selected. This is variably called a random subspace method, attribute bagging, or feature bagging. By convention, specifically, a SRS with replacement is used. For each selected feature, a decision tree is estimated using the entire set of observations.
Generally, very deep trees are allowed to be built, because the random selection of features across iterations should correct for overfitting.