On the asymptotics of random forests
Author:
Preprint
Abstract
The last decade has witnessed a growing interest in random forest models which are recognized to exhibit good practical performance, especially in high-dimensional settings. On the theoretical side, however, their predictive power remains largely unexplained, thereby creating a gap between theory and practice. The aim of this paper is twofold. Firstly, we provide theoretical guarantees to link finite forests used in practice (with a finite number M of trees) to their asymptotic counterparts. Using empirical process theory, we prove a uniform central limit theorem for a large class of random forest estimates, which holds in particular for Breiman's original forests. Secondly, we show that infinite forest consistency implies finite forest consistency and thus, we state the consistency of several infinite forests. In particular, we prove that q quantile forests---close in spirit to Breiman's forests but easier to study---are able to combine inconsistent trees to obtain a final consistent prediction, thus highlighting the benefits of random forests compared to single trees.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Consistency of random forestsThe Annals of Statistics, 2015
- New insights into Approximate Bayesian ComputationAnnales de l'Institut Henri Poincaré, Probabilités et Statistiques, 2015
- Consistency of random survival forestsStatistics & Probability Letters, 2010
- quantregForest: Quantile Regression Forests2006
- Gene selection and classification of microarray data using random forestBMC Bioinformatics, 2006
- A Distribution-Free Theory of Nonparametric Regression2002
- The random subspace method for constructing decision forestsIEEE Transactions on Pattern Analysis and Machine Intelligence, 1998
- A Probabilistic Theory of Pattern Recognition1996