A random forest guided tour

The random forest algorithm, proposed by L. Breiman in 2001, has been extremely successful as a general-purpose classification and regression method. The approach combines several randomized decision trees and aggregates their predictions by averaging, showing excellent performance in settings where the number of variables is much larger than the number of observations. It is versatile enough to be applied to large-scale problems, is easily adapted to various ad hoc learning tasks, and returns measures of variable importance. This article surveys the method's landscape rather than proposing a new algorithm.

The review covers the most recent theoretical and methodological developments for random forests, placing emphasis on the mathematical forces driving the algorithm, with special attention given to the selection of parameters, the resampling mechanism, and variable importance measures. It is explicitly intended to provide non-experts easy access to the main ideas, making it a valuable tutorial-style reference that consolidates scattered theoretical results on why and how random forests work.