compress.matrix: Compress a matrix using the truncated SVD

In awllee/SC1ExamplePackage: Compress Images Using the Truncated SVD

Description

Compress a matrix using the truncated SVD. Exactly one of rank, ratio and avg.error should be specified. The compression ratio will be at least the given value, or the average error will be at most the given value. The average error is the Frobenius norm of the error divided by the square root of the number of elements in the matrix.

Usage

compress.matrix(mtx, rank = NA, ratio = NA, avg.error = NA)

Arguments

mtx	matrix
rank	rank of the truncated SVD
ratio	compression ratio
avg.error	average error

Value

compressed.matrix object with decomposition, rank, error, avg.error

Examples

compress.matrix(matrix(1:12, 4, 3), rank=2)
compress.matrix(matrix(1:12, 4, 3), ratio=1.1)
compress.matrix(matrix(1:12, 4, 3), avg.error=0.5)

awllee/SC1ExamplePackage documentation built on Sept. 1, 2021, 4:06 a.m.