sparseLRMatrix provides a single matrix S4 class called
sparseLRMatrix which represents matrices that can be expressed as the sum of sparse matrix and a low rank matrix. We also provide an efficient SVD method for these matrices by wrapping the
RSpectra SVD implementation.
Eventually, we will fully subclass
Matrix::Matrix objects, but the current implementation is extremely minimal.
You can install the released version of sparseLRMatrix from CRAN with:
You can install the development version with:
# install.packages("remotes") remotes::install_github("RoheLab/sparseLRMatrix")
library(sparseLRMatrix) #> Loading required package: Matrix library(RSpectra) set.seed(528491) n <- 50 m <- 40 k <- 3 A <- rsparsematrix(n, m, 0.1) U <- Matrix(rnorm(n * k), nrow = n, ncol = k) V <- Matrix(rnorm(m * k), nrow = m, ncol = k) # construct the matrix, which represents A + U %*% t(V) X <- sparseLRMatrix(sparse = A, U = U, V = V) s <- svds(X, 5) # efficient
And a quick sanity check