RcppPlanc (version 2.0.13)

bppnnls: Block Principal Pivoted Non-Negative Least Squares

Description

Use the BPP algorithm to get the nonnegative least squares solution. Regular NNLS problem is described as optimizing \(\min_{x\ge0}||CX - B||_F^2\) where \(C\) and \(B\) are given and \(X\) is to be solved. bppnnls takes \(C\) and \(B\) as input. bppnnls_prod takes \(C^\mathsf{T}C\) and \(C^\mathsf{T}B\) as input to directly go for the intermediate step of BPP algorithm. This can be useful when the dimensionality of \(C\) and \(B\) is large while pre-calculating \(C^\mathsf{T}C\) and \(C^\mathsf{T}B\) is cheap.

Usage

bppnnls(C, B, nCores = 2L)
bppnnls_prod(CtC, CtB, nCores = 2L)

Value

The calculated solution matrix in dense form.

Arguments

C

Input dense \(C\) matrix

B

Input \(B\) matrix of either dense or sparse form

nCores

The number of parallel tasks that will be spawned. Default 2

CtC

The \(C^\mathsf{T}C\) matrix, see description.

CtB

The \(C^\mathsf{T}B\) matrix, see description.

Examples

set.seed(1)
C <- matrix(rnorm(250), nrow = 25)
B <- matrix(rnorm(375), nrow = 25)
res1 <- bppnnls(C, B)
dim(res1)
res2 <- bppnnls_prod(t(C) %*% C, t(C) %*% B)
all.equal(res1, res2)