The Holdout Randomization Test for Feature Selection in Black Box Models

DOI10.1080/10618600.2021.1923520OpenAlexW3175769723MaRDI QIDQ5083362

Haoran Zhang, Raúl Rabadán, Wesley Tansey, Victor Veitch, David M. Blei

Publication date: 22 June 2022

Published in: Journal of Computational and Graphical Statistics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1811.00645

zbMATH Keywords

feature selectionconditional independence testingblack box models

Mathematics Subject Classification ID

Statistics (62-XX)

Related Items (6)

pyhrt ⋮ Understanding complex predictive models with ghost variables ⋮ Learning to increase the power of conditional randomization tests ⋮ Dimension-agnostic inference using cross U-statistics ⋮ Unnamed Item ⋮ On the power of conditional independence testing under model-X

Uses Software

• QuickMMCTest
• conformalInference
• GeneralisedCovarianceMeasure
• functionalmf
• DeepPINK
• KnockoffGAN

Cites Work

• Unnamed Item
• Distribution-Free Predictive Inference For Regression
• The Hardness of Conditional Independence Testing and the Generalised Covariance Measure
• High-dimensional variable selection
• Statistical modeling: The two cultures. (With comments and a rejoinder).
• The control of the false discovery rate in multiple testing under dependency.
• QuickMMCTest: quick multiple Monte Carlo testing
• p-Values for High-Dimensional Regression
• Bayesian Approximate Kernel Regression With Variable Selection
• Panning for Gold: ‘Model-X’ Knockoffs for High Dimensional Controlled Variable Selection
• Iterative random forests to discover predictive and stable high-order interactions
• The Conditional Permutation Test for Independence While Controlling for Confounders
• Deep Knockoffs
• Gene hunting with hidden Markov model knockoffs

This page was built for publication: The Holdout Randomization Test for Feature Selection in Black Box Models