Importance Sampled Learning Ensembles

By Jerome H. Friedman, Bogdan E. Popescu. 2003

134 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.

Citations in the corpus

XGBoost: A Scalable Tree Boosting System (2016, background): “In order to do so eﬃciently, the algorithm must ﬁrst sort the data according to feature values and visit the data in sorted order to accumulate the gradient statistics for the structure score in Eq (7).”
Learning Nonlinear Functions Using Regularized Greedy Forest (2011, background): “Although [18] discusses various techniques regarding rules, we focus on the aspect of the two-stage approach which [18] derives from [17], since it is the most relevant portion to our work due to its contrast with our interleaving approach.”
A high-bias, low-variance introduction to Machine Learning for physicists (2018, unknown): “In practice, additional regularization such as shrinkage(Friedman, 2002) and feature subsampling(Breiman, 2001; Friedman et al., 2003) is also used.”
A statistical approach to rule learning (2006, background): “(2001) and RuleFit by Friedman and Popescu (2005).”
pre: An R Package for Fitting Prediction Rule Ensembles (2017, methodology): “Based on the ﬁndings of Friedman (2001) and Friedman and Popescu (2003), who found small nonzero values of the learning rate to perform best for ensembles of small decision trees, the learning rate in pre is set to .01 by default.”
Interpretable Random Forests via Rule Extraction (2020, methodology): “Both methods also extract large collection of rules from tree ensembles: RuleFit uses Importance Sampled Learning Ensemble (ISLE, Friedman & Popescu, 2003) Node harvest is based on random forests.”
ENDER: a statistical framework for boosting decision rules (2010, background): “That is why the procedure for finding m works on a subsample of original data, that is a fraction ζ of all training examples, drawn without replacement (Friedman and Popescu 2003).”
From Predictive to Prescriptive Analytics (2014, methodology): “th for a single sample path of an evolving market with each method trained on the rst 1,000 non-IID points. Note the logarithmic horizontal axis on the left and vertical axis on the right. processing [17] (see also x16.3 of [39]). These also require an appropriate measure of loss; we discuss this in”
A survey on missing data in machine learning (2021, methodology): “Ensemble Generation The general ensemble algorithm creation which was formalized by [112] consists of two steps as stated above.”