Panning for Gold: Model-free Knockoffs for High-dimensional Controlled Variable Selection

By E.J. Candes, Yingying Fan, Lucas Janson, Jinchi Lv. 2016

304 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/a786a7c6a9cdcc21dab30e1cd4f23169c086de74

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

A modern maximum-likelihood theory for high-dimensional logistic regression (2018, background): “Naturally, using classical predictions would lead to grossly incorrect p-values and confidence statements, a major issue first noticed in [11].”
Bootstrapping and sample splitting for high-dimensional, assumption-lean inference (2016, background): “While this is a larger literature than can be addressed completely here, it includes early work on model selection [35] and model averaging interpretations [31]; the impossibility results of [38] and [16] on random X and model misspecification; methods based on resampling or sample splitting [18, 19”
The conditional permutation test (2018, methodology): “While Theorem 4 provides an upper bound on the Type I error for both the CPT and the CRT, we do not yet have a comparison between the two.”
A Critical Review of LASSO and Its Derivatives for Variable Selection Under Dependence Among Covariates (2020, methodology): “Furthermore, more recent techniques like the Knockoﬀ ﬁlter (Barber et al. (2015), Candes et al. (2018)) or SLOPE (Bogdan et al. (2015)) have been introduced to control some measures of the type I error.”
Covariate powered cross‐weighted multiple testing (2017, methodology): “If avoiding dependence by modifying the analysis upstream of the multiple testing treatment is not possible, the analyst should also consider whether marginal hypothesis tests are indeed more appropriate than, say, a multivariate model with FDR guarantees (Candès et al., 2018; Sesia et al., 2017).”
Splitting strategies for post-selection inference (2021, methodology): “Cox (1975) analysed data splitting in a simple inferential problem involving many normal means and found it to be competitive against a natural alternative.”
The conditional permutation test for independence while controlling for confounders (2018, methodology): “The ‘model-X knockoffs’ framework of Candès et al. (2018) also extends the CRT technique to the high dimensional variable selection setting, where each of p features is tested in turn for conditional independence with the response Y , with the goal of false discovery rate control.”
False discovery rate control with e‐values (2020, methodology): “Examples include the permutation test, the conditional randomization test [5] and conformal prediction [15].”
Please Stop Permuting Features: An Explanation and Alternatives (2019, background): “(2015) and Candes et al. (2018), although the latter somewhat more stringent conditions. Similar ideas also appear in Tuv et al. (2009) and also (more specific to the context of linear regression) in Wu et al.”
Getting Better from Worse: Augmented Bagging and a Cautionary Tale of Variable Importance (2020, background): “In a similar fashion to knockoffs (Barber et al., 2015; Candes et al., 2018), such noise features are thus independent of the response Y given X.”