CVEQUAL: Cross-validation function for EQUAL

In cescwang85/EQUAL: An efficient ADMM algorithm for high dimensional precision matrix estimation via penalized quadratic loss

Cross-validation function for EQUAL

Description

Cross-validation function for EQUAL

Usage

cvEQUAL(
  X,
  K = 5,
  type = TRUE,
  sdiag = FALSE,
  lambda = NULL,
  lambda.min = sqrt(log(ncol(X))/nrow(X)),
  nlambda = 50,
  err = 10^(-5),
  maxIter = 1000,
  rho = 1
)

Arguments

X	data matrix of dimension n*p.
K	the number of folds. Default is 5.
type	Should the loss function be symmetric? Default is TRUE.
sdiag	Should diagonal of inverse covariance be penalized? Default is FALSE.
lambda	user supplied tuning parameter; Default is NULL and the program compute its own sequence based on nlambda.
lambda.min	smallest value for lambda, as a fraction of lambda.max which is available when lambda is NULL. Default is sqrt(log(p)/n).
nlambda	the length of the tuning parameter sequence which is available when lambda is NULL. Default is 50.
err	the precision used to stop the convergence. Default is 1e-5. Iterations stop when average absolute parameter change is less than err.
maxIter	Maximum number of iterations. Default is 1000.
rho	step parameter for the ADMM. Default is 1.

Value

A list with components

Omega	the estimated p*p precision matrix.
cvlambda	the chosen lambda by cross-validation.
lambda	the used lambda list for cross-validation.
cvloss	the empirical loss of cross-validation related to lambda.

cescwang85/EQUAL documentation built on Nov. 26, 2022, 1:27 a.m.