Unlock: Random Forests

Random forests combine bagging with random feature subsampling to decorrelate trees, reducing ensemble variance beyond what pure bagging achieves. Out-of-bag estimation, variable importance, consistency theory, and practical strengths and weaknesses.