Knockoffs (statistics)
≈ 1 min read
Names in other languages
No names in other languages are held for this entry yet.
Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.
In statistics, the knockoff filter, or simply knockoffs, is a framework for variable selection. It was originally introduced for linear regression by Rina Barber and Emmanuel Candès, and later generalized to other regression models in the random design setting. Knockoffs has found application in many practical areas, notably in genome-wide association studies.
In statistics, the knockoff filter, or simply knockoffs, is a framework for variable selection. It was originally introduced for linear regression by Rina Barber and Emmanuel Candès, and later generalized to other regression models in the random design setting. Knockoffs has found application in many practical areas, notably in genome-wide association studies. Consider a linear regression model with response vector \mathbf y and feature matrix \mathbf X , which is treated as deterministic.
A matrix \tilde{\mathbf X} is said to be knockoffs of \mathbf X if it does not depend on \mathbf y and satisfies \mathbf X_i^\top\mathbf X_j=\mathbf X_i^\top\tilde{\mathbf X}_j=\tilde{\mathbf X}_i^\top\mathbf X_j=\tilde{\mathbf X}_i^\top\tilde{\mathbf X}_j for i\ne j . Barber and Candès showed that, equipped with a suitable feature importance statistic, fixed-X knockoffs can be used for variable selection while controlling the false discovery rate (FDR).
Text derived from https://en.wikipedia.org/w/index.php?oldid=1087053526 under CC BY-SA 4.0; modifications have been made.