Grace_multi_knockoff: Function to conduct Grace-AKO and Grace

In mxxptian/GraceAKO: Grace-AKO: A Novel and Stable Knockoff Filter for Variable Selection Incorporating Gene Network Structures

Function to conduct Grace-AKO and Grace

Description

Function to conduct Grace-AKO and Grace

Usage

Grace_multi_knockoff(
  X,
  Y,
  L_1,
  L,
  fdr,
  lambda.L,
  lambda.1,
  lambda.2,
  n_bootstraps = 25,
  gamma = 0.3,
  gamma_min = 0.1
)

Arguments

X	standardized n (number of rows) by p (number of columns) design matrix.
Y	centered n by 1 vector of the response variable.
L_1	p by p (weighted) processed matrix of a graph for Grace-AKO.
L	p by p (weighted) adjacency matrix of a graph for Grace.
fdr	the pre-specified false discovery rate level
lambda.L	tuning parameters of the penalty weight matrix
lambda.1	tuning parameters of the L_1 penalty
lambda.2	tuning parameters of the ridge penalty
n_bootstraps	the number of multiple knockoffs
gamma	the pre-specified percentile value used for fixed aggregation
gamma_min	the pre-specified percentile value used for adaptive aggregation

mxxptian/GraceAKO documentation built on Oct. 5, 2022, 11:17 p.m.