The Estimation of Prediction Error

By Bradley Efron. 2004

348 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/f5d74edd1b4023930145d756c73ae4e48165c930

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

A robust coefficient of determination for regression (2010, methodology): “Efron (2004) shows the link between covariance penalty methods and cross-validation and related nonparametric bootstrap techniques.”
The composite absolute penalties family for grouped and hierarchical variable selection (2009, background): “While the CAP estimates are not sparser than LASSO estimates in the number of variables sense, they result in more parsimonious use of degrees of freedom and more stable estimates [ 7 ].”
Optimal Model Assessment, Selection, and Combination (2006, methodology): “Efron (2004) showed that the covariance penalty is more accurate than cross-validation.”
‘All models are wrong...’: an introduction to model uncertainty (2012, background): “Important developments in general covariance penalty theory (Friedman, Hastie and Tibshirani, 2001; Efron, 2004) showed that the degrees of freedom can © 2012 The Authors.”
Prediction error estimation: a comparison of resampling methods (2005, methodology): “In two recent assessments of resampling techniques for error estimation (Braga-Neto and Dougherty, 2004; Efron, 2004), feature selection was not included as part of the resampling procedures, causing the conclusions to be inappropriate for the high-dimensional setting.”
Estimation of prediction error by using K-fold cross-validation (2011, methodology): “Although leave-oneout cross-validation is studied in many researches (for example, Stone 1974, 1977; Efron 2004), K-fold cross-validation may be preferred from a computational standpoint.”
Spatial weights matrix selection and model averaging for spatial autoregressive models (2017, methodology): “To interpret the second term of Cs, following Efron (2004) and Zou et al. (2007), we define the degrees of freedom of the sth model as dfs = cov(µ̂Ts , yT), where cov(a, b) = E[{a − E(a)}{b − E(b)}T] for two random vectors a and b.”
Akaike's Information Criterion, Cp and Estimators of Loss for Elliptically Symmetric Distributions (2014, result): “Note that there is also an equivalence with other model selection criteria, such as those investigated in Li (1985), Shao (1997) and Efron (2004).”
Testing for Homogeneity with Kernel Fisher Discriminant Analysis (2007, methodology): “The recentering parameter d 1 (Σ 1 , γ n ) in our procedure coincides with the so-called degrees of freedom in smoothing spline regression, which were often advocated to provide a relevant measure of complexity for model selection (see Efron, 2004).”
A point process framework for relating neural spiking activity to spiking history, neural ensemble, and extrinsic covariate effects. (2005, methodology): “Additionally, an equivalence between AIC and cross-validation in model comparison problems has been established (Stone 1977), and model-based penalty methods (e.g., AIC) outperform crossvalidation in important aspects when assessing prediction performance (Efron 2004).”