Decorrelation projection + eclairs — whiten • decorrelate

Efficient decorrelation projection using eclairs decomposition

Usage

whiten(X, k = ncol(X), lambda = NULL)

Arguments

X: matrix to be transformed so *columns* are independent
k: the rank of the low rank component
lambda: specify lambda and override value estimated by eclairs()

Value

data rotated and scaled according to the regularized sample covariance of the input data

Examples

library(Rfast)

n <- 800 # number of samples
p <- 200 # number of features

# create correlation matrix
Sigma <- autocorr.mat(p, .9)

# draw data from correlation matrix Sigma
Y <- rmvnorm(n, rep(0, p), sigma = Sigma * 5.1, seed = 1)

# eclairs decomposition
ecl <- eclairs(Y)

# whitened Y
Y.transform <- decorrelate(Y, ecl)

# Combine eclairs and decorrelate into one step
Y.transform2 <- whiten(Y)