Controlling the False Discovery Rate via knockoffs: is the +1 needed?

04/28/2022

∙

Barber and Candès (2015) control of the FDR in feature selection relies on estimating the FDR by the number of knockoff wins +1 divided by the number of original wins. We show that the +1 is necessary in typical scenarios.

READ FULL TEXT

Controlling the False Discovery Rate via knockoffs: is the +1 needed?

Stepdown SLOPE for Controlled Feature Selection

ADAGES: adaptive aggregation with stability for distributed feature selection

Feature Selection for Machine Learning Algorithms that Bounds False Positive Rate

Controlling the False Discovery Rate in Structural Sparsity: Split Knockoffs

Estimating and Controlling the False Discovery Rate for the PC Algorithm Using Edge-Specific P-Values

CoxKnockoff: Controlled Feature Selection for the Cox Model Using Knockoffs

ECKO: Ensemble of Clustered Knockoffs for multivariate inference on fMRI data

Controlling the False Discovery Rate via knockoffs: is the +1 needed?

Related Research

Stepdown SLOPE for Controlled Feature Selection

ADAGES: adaptive aggregation with stability for distributed feature selection

Feature Selection for Machine Learning Algorithms that Bounds False Positive Rate

Controlling the False Discovery Rate in Structural Sparsity: Split Knockoffs

Estimating and Controlling the False Discovery Rate for the PC Algorithm Using Edge-Specific P-Values

CoxKnockoff: Controlled Feature Selection for the Cox Model Using Knockoffs

ECKO: Ensemble of Clustered Knockoffs for multivariate inference on fMRI data