Centering a kernel matrix

It is equivalent to compute `K` over centered data (i.e. the mean of each column is subtracted) in Feature Space.

Usage

centerK(K)

Arguments

K

Kernel matrix (class "matrix").

Value

Centered `K` (class "matrix").

Examples

dat <- matrix(rnorm(250),ncol=50,nrow=5)
K <- Linear(dat)
centerK(K)
#>            [,1]       [,2]       [,3]       [,4]       [,5]
#> [1,]  36.683262 -13.383489 -12.032447  -9.653916  -1.613410
#> [2,] -13.383489  45.009171 -12.265942 -11.884602  -7.475137
#> [3,] -12.032447 -12.265942  37.211863  -4.253565  -8.659909
#> [4,]  -9.653916 -11.884602  -4.253565  43.311665 -17.519581
#> [5,]  -1.613410  -7.475137  -8.659909 -17.519581  35.268038