computeProjectionFromKernel: computeProjectionFromKernel

In semisupKernelPCA: Kernel PCA projection, and semi-supervised variant.

Description

Compute the kernel PCA projection from an input kernel matrix.

Usage

computeProjectionFromKernel(kernel, dims=2, eigentype=c("basic", "irlba"))

Arguments

kernel	n x n kernel matrix
dims	number of output dimensions for the projection. Cannot exceed n.
eigentype	Indicates the eigendecomposition routine that should be used, either the standard ("basic"), or an optimization for the extraction of a few major eigenpairs ("irlba")

Value

n x dims matrix of the projected data.

Author(s)

Pierrick Bruneau

References

Bishop, C. M. (2006) Pattern recognition and machine learning. Springer.

See Also

computeStandardKernel

Examples

data(iris)
irisdat <- as.matrix(iris[,1:4])
kernel <- computeKernel(irisdat)
proj <- computeProjectionFromKernel(kernel)

semisupKernelPCA documentation built on May 29, 2017, 8:59 p.m.