denoise_rectangular_matrix {GSSTDA}

Rectangular Matrix Denoiser.

Description

It takes a rectangular matrix composed by the addition of a signal matrix and a Gaussian noise matrix and returns a matrix of the same dimension that is denoised through a Singular Value Decomposition truncation process. The selection of the number of singular values is chosen following the proposal by "The optimal hard threshold for singular values is \sqrt(4/ 3)". It should be used after the function flatten_normal_tiss.

Usage

denoise_rectangular_matrix(matrix_flatten_normal_tiss)

Arguments

matrix_flatten_normal_tiss

A rectangular noisy matrix to denoise. It is return by flatten_normal_tiss function.

Value

A the normal space which has the same dimension denoised version of the matrix returned by flatten_normal_tiss.

Examples

denoise_rectangular_matrix(matrix(c(1,2,3,4,5,2,3,1,2,3),ncol = 2))