Skip to contents

Project new data onto SVD left singular vectors

Source: R/svd-factor.R
svd_project.Rd

Computes t(U) %*% Y, where U is the matrix of left singular vectors stored in factor. Routes through a GPU backend when the factor was computed in "fast" precision and a device copy of t(U) is available.

Usage

svd_project(factor, Y)

Arguments

factor

An amSVD object from svd_factor.

Y

Numeric matrix or vector with nrow(Y) equal to the number of rows of the original source matrix.

Value

Numeric matrix of dimensions k x ncol(Y) containing the projected coordinates.

Examples

m <- matrix(rnorm(30), nrow = 6)
A <- adgeMatrix(m)
fac <- svd_factor(A, k = 2L)
coords <- svd_project(fac, m)