Transform matrix to use Mahalanobis distance instead of Euclidean one.

maha_trans(U, estim = covrob_ogk(U))

Arguments

U

A matrix (e.g. PC scores).

estim

List of location and scatter estimates, $cov and $center.

Value

U, transformed.

Examples

X <- readRDS(system.file("testdata", "three-pops.rds", package = "bigutilsr"))
svd <- svds(scale(X), k = 5)

U <- svd$u
dist1 <- dist_ogk(U)

U.maha <- maha_trans(U)
dist2 <- rowSums(U.maha^2)
all.equal(dist2, dist1)
#> [1] TRUE