Random forest is a machine learning ensemble method based on choosing random subsets of observations and variables for each of many decision trees.
Overview
Random forests are an ensemble learning method for classification or regression that operate by constructing a multitude of decision trees at training time and aggregating their outputs: in regression by averaging, in classification either by averaging probabilities or by majority vote. The randomness in the decision trees follows from (a) using a new bootstrapped version of the original sample for each tree and (b) using a random subsample of the explanatory variables at each node of each tree.
References
- Random Forest page maintained by Leo Breiman and Adele Cutler, the creators of the algorithm.
- Wikipedia pages on Random Trees, Random Forest and Ensemble Learning.
