On bagging and nonlinear estimation

By Jerome H. Friedman, Peter Hall. 2007

215 citations recorded by the source corpus.

How researchers see this paper

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

In Statistics and Probability

Citations in the corpus

PREDICTIVE LEARNING VIA RULE ENSEMBLES (2008, methodology): “A “bagged” ensemble [Breiman (1996)] is obtained by using squared-error loss, L ( y, ˆ y ) = ( y − ˆ y ) 2 , and setting ν = 0, and η = N/ 2 or, equivalently, choosing S m (line 3) to be a bootstrap sample [Friedman and Hall (2007)].”
Offline Data-Driven Evolutionary Optimization Using Selective Surrogate Ensembles (2019, methodology): “To this end, a large number of surrogates are generated ofﬂine using the bagging technique [39], [40] and a subset of them is adaptively selected for ﬁtness estimation as the evolutionary optimization process proceeds.”
Multiple classifier systems for robust classifier design in adversarial environments (2010, background): “These techniques, and bagging in particular, have been extensively studied in the literature, in terms of why and under which conditions they are able to improve classification accuracy with respect to an individual base classifier, or the approximation error in regression problems, for bagging [9,1”
Inferences from Cross-Sectional, Stochastic Frontier Models (2009, methodology): “Theoretical properties of bagging predictors have been examined by Friedman and Hall (1999), Bühlmann and Yu (2002), and Fushiki et al. (2005).”
Inverse random under sampling for class imbalance problem and its application to multi-label classification (2012, background): “by making the majority (negative) class subsets have fewer examples than the minority (positive) class; then, control the false positive rate (fpr) by classifier bagging [24,29,28] i.”
Confidence Intervals for Random Forests: The Jackknife and the Infinitesimal Jackknife (2013, background): “Jaeckel (1972) originally introduced the infinitesimal jackknife as a practical approximation to the first-order variance of an estimator (in our case, the righthand side of (21)).”
Random Forests and Adaptive Nearest Neighbors (2006, background): “That is, the expected number of k -PNN’s is of order k (log n ) d − 1 .”
Out-of-bag estimation of the optimal sample size in bagging (2010, methodology): “A common choice for the sample size in subagging is mwor = n/2 [11].”
Bias in random forest variable importance measures: Illustrations, sources and a solution (2007, methodology): “5 as suggested by Friedman and Hall [30].”
Leveraging Bagging for Evolving Data Streams (2010, methodology): “Friedman and Hall [15] used an asymptotic truncated Taylor series of the estimate to show that in the limit of infinite samples, bagging reduces the variance of non-linear components.”
Asymptotic Theory for Random Forests (2014, background): “As observed by Buja and Stuetzle [10], Chen and Hall [11], Friedman and Hall [17] and others, estimators of this form can exhibit surprising properties even in simple situations.”
OBSERVATIONS ON BAGGING (2006, background): “The scenarios are adapted from Friedman and Hall (2000). Scenario N p X σ2 K f(x) 1 800 1 U [0, 1] 1 2 I(x > 0.”