The Elements of Statistical Learning: Data Mining, Inference, and Prediction

By Trevor J. Hastie, Jerome H. Friedman, Robert Tibshirani. 2001

22,053 citations recorded by the source corpus.

Subfields: Information Systems, Statistics and Probability, Statistical Learning Theory

How researchers see this paper

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

In Information Systems

In Statistics and Probability

In Statistical Learning Theory

Citations in the corpus

Deep Clustering for Unsupervised Learning of Visual Features (2018, methodology): “Unsupervised learning has been widely studied in the Machine Learning community [12], and algorithms for clustering, dimensionality reduction or density ar X iv :1 80 7.”
The Loss Surfaces of Multilayer Networks (2014, background): “Also note that statistical learning theory heavily relies on this assumption [Hastie et al., 2001] even when the model under consideration is much simpler than a neural network.”
A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, positive): “According to the learning theory for supervised learning [48], the performance of an old task is improved with replaying more old training samples that approximate its distribution, but resulting in potential privacy issues and a linear increase in resource overhead .”
Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions (2018, neutral): “The two margins of the confusion matrix are distinguished by Barocas et al. as [10]: (also known as error rate balance [17], equalized odds [53], or conditional procedure accuracy equality [13]) this is also known as conditional use accuracy equality [13].”
Distributed delayed stochastic optimization (2011, methodology): “] ≤ σ2 holds.1 Several commonly used functions satisfy the above assumptions, for example: (i) The logistic loss: F(x;ξ) = log[1+exp(hx,ξi)], the objective for logistic regression in statistics (e.g. [HTF01]). The objective Fsatisﬁes Assumptions A and B so long as kξk is bounded. (ii) Least squares ”
Deep learning in neural networks: An overview (2014, background): “tion Maximization (EM) (Dempster et al., 1977; Friedman et al., 2001), e.g., (Bottou, 1991; Bengio, 1991; Bourlard and Morgan, 1994; Baldi and Chauvin, 1996; Jordan and Sejnowski, 2001; Bishop, 2006; Hastie et al., 2009; Poon and Domingos, 2011; Dahl et al., 2012; Hinton et al., 2012a; Wu and Shao, ”
On the Art and Science of Machine Learning Explanations (2018, methodology): “on (ICE) Plots Partial dependence (PD) plots are a widely-used method for describing the average predictions of a complex model g across some partition of data Xfor some interesting input feature X j [10]. Individual conditional expectation (ICE) plots are a newer method that describes the local beh”
Consistent Individualized Feature Attribution for Tree Ensembles (2018, methodology): “ce introduced by Breiman et al. in 1984 [3] is based on gain. Gain is the total reduction of loss or impurity contributed by all splits for a given feature. Though its motivation is largely heuristic [11], gain is widely used as the basis for feature selection methods [5, 13, 25]. (2) Split Count: A”
Visualizing the effects of predictor variables in black box supervised learning models (2016, background): “Moreover, for f ( x ) = f J ( x J ) f \ J ( x \ J ), PD plots have multiplicative recovery (up to a multiplicative constant) even when X J and X \ J are dependent (Hastie and Friedman (2009)).”
Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation (2013, background): “ors x 1;:::x N are increasingly sparse in the feature space X. A consequence of this curse of dimensionality is that for many x 2X, f^(x) represents an extrapolation rather than an interpolation (see Hastie et al., 2009 for a more complete discussion). Extrapolation may be of particular concern when”
Fairness in Criminal Justice Risk Assessments: The State of the Art (2017, methodology): “Discussions in statistics about generalization error (Hastie et al. 2009:section 7.2) can provide one set of terms whereas concerns about errors from statistical tests can provide another.”
Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems (2019, background): “that is known to result in disentangled and interpretable low-dimensional representations of the observed data [58].”