EFM (version 1.3.0)

robust_perturb_pca: Random Perturbation Robust PCA

Description

Robust PCA by adding random symmetric perturbation to covariance matrix.

Usage

robust_perturb_pca(data, m, epsilon = 0.001, sigma = 0.01)

Value

List with Ap, Dp, lambda.

Arguments

data

T x N matrix.

m

Components.

epsilon

Perturbation scale.

sigma

Noise sd.

Examples

if (FALSE) {
set.seed(456)
N <- 80; T <- 100; m.true <- 2
X <- matrix(rnorm(T * N), T, N)
rpca.out <- EFM::robust_perturb_pca(X, m = m.true)
print(round(rpca.out$Ap[1:5, 1:2], 3))
}