HM_proj: HM-lasso algorithm

In celiaescribe/BDcocolasso: Implementation of CoCoLasso and Block Descent CoCoLasso

Description

Finds the nearest positive semi-definite matrix with respect to the Frobenius norm

Usage

HM_proj(
  sigmaHat,
  R = NULL,
  a = 1,
  iter_max = 1000,
  epsilon = 1e-04,
  mu = 10,
  tolerance = 1e-04,
  norm = "F"
)

Arguments

sigmaHat	covariance matrix
R	weight matrix
a	hmlasso (1) or cocolasso (0)
iter_max	Number maximum of iterations
epsilon	Stopping criterion
mu	Penalty parameter of ADMM algorithm
tolerance	Tolerance parameter for the convergence of primal and dual residual
norm	Frobenius or max norm

Value

list containing

• Ak Projected matrix

See Also

https://arxiv.org/pdf/1811.00255.pdf

Examples

 
M = matrix(-1,20,20)
mat_proj <- BDcocolasso::HM_proj(sigmaHat=M)

celiaescribe/BDcocolasso documentation built on Feb. 11, 2020, 11:41 p.m.