Random Forests are a bread and butter model/algorithm for machine learning. They were first described in their current form as recently as 2001 in a classic paper by the late Leo Breiman. Even with the rise in popularity of artificial neural networks, they find practical use in a variety of situations.

There are a wealth of resources detailing how Random Forests work. This article will very briefly review them before turning to the main focus: how to fit them faster with warm starts and out-of-bag cross-validation. With these two techniques, hyper-parameters selection can be sped up substantially, reducing fitting time.

#editors-pick #artificial-intelligence #machine-learning #data-science #towards-data-science