Standard errors for bagged and random forest estimators

By Joseph Sexton, Petter Laake. 2009

119 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Local Linear Forests (2018, methodology): “Early versions of confidence intervals for random forests, backed by heuristic arguments and empirical evidence, were proposed by Sexton and Laake (2009) and Wager, Hastie, and Efron (2014).”
Generalized random forests (2016, methodology): “Fig 2: Comparison of quantile regression using generalized random forests and the quantregForest package of Meinshausen (2006). In both cases, we have n = 2, 000 independent and identically distributed examples where Xi is uniformly distributed over [−1, 1]p with p = 40, and Yi is Gaussian condition”
Auditing Black-Box Models Using Transparent Model Distillation With Side Information (2017, methodology): “We employ a bootstrap-of-little-bags approach originally developed for bagged models in [39] to obtain pointwise confidence intervals for GA2M feature contributions, and the difference of two GA2M models’ feature contributions.”
Estimation and Accuracy after Model Selection (2013, background): “A particularly apt reference is Sexton and Laake (2009), who also provide modified bootstrap estimates of accuracy for estimators involving model selection.”
Asymptotic Theory for Random Forests (2014, result): “Our work builds on recent contributions by Duan [14], Sexton and Laake [29] and Wager et al.”
Quantifying Uncertainty in Random Forests via Confidence Intervals and Hypothesis Tests (2014, background): “Sexton and Laake (2009) propose estimating the standard error of bagged trees and random forests using jackknife and bootstrap estimators.”
Getting Better from Worse: Augmented Bagging and a Cautionary Tale of Variable Importance (2020, background): “…normality (Mentch and Hooker, 2016; Wager and Athey, 2018), and rates of convergence (Peng et al., 2019) as well as means by which standard errors (Sexton and Laake, 2009), conﬁdence intervals (Wager et al., 2014; Mentch and Hooker, 2016), and hypothesis testing procedures (Mentch and Hooker,…”
Ensemble Trees and CLTs: Statistical Inference for Supervised Learning (2014, unknown): “Sexton and Laake [2009] propose estimating the standard error of bagged trees and random forests us-ing jackknife and bootstrap estimators.”
Confidence Intervals for Random Forests: The Jackknife and the Infinitesimal Jackknife (2013, methodology): “Other approaches that rely on forming second-order bootstrap replicates have been studied by Duan (2011) and Sexton and Laake (2009).”
Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges (2021, background): “However, more sophisticated alternatives exist to provide unbiased estimates (Sexton & Laake, 2009).”