This package implements the Knockoff Filter, which is a powerful and versatile tool for controlled variable selection.
The procedure is based on the contruction of artificial 'knockoff copies' of the variables present in the given statistical model. Then, it selects those variables that are clearly better than their corresponding knockoffs, based on some measure of variable importance. A wide range of statistics and machine learning tools can be exploited to estimate the importance of each variable, while guaranteeing finite-sample control of the false discovery rate (FDR).
The Knockoff Filter controls the FDR in either of two statistical scenarios:
The "model-X" scenario: the response Y can depend on the variables X=(X_1,…,X_p) in an arbitrary and unknown fashion, but the distribution of X must be known. In thise case there are no constraints on the dimensions n and p of the problem.
The "fixed-X" scenario: the response Y depends upon X through a homoscedastic Gaussian linear model and the problem is low-dimensional (n ≥q p). In this case, no modeling assumptions on X are required.
For more information, see the website below and the accompanying paper.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.