matfaust::proj::spcol Class Reference

Functor for the SPCOL projector. More...

Public Member Functions
function	spcol (shape, k, varargin)
Public Member Functions inherited from matfaust::proj::proj_gen
function	subsref (self, S)

Additional Inherited Members
Protected Attributes inherited from matfaust::proj::proj_gen
Property	constraint

Detailed Description

Functor for the SPCOL projector.

A, the image matrix, is defined by \( \forall j \in \{1,\ldots,shape(2)\} \) the j-th column \( A_{*,j} \) is such that \( \|A_{*,j}\|_0 = k, \| A\|_F = 1 \) (if normalized is true).

Usage

    spcol(shape, k): returns a SPCOL projector (functor), shape defines the size of the input matrix (e.g. [1, 10]), k defines the sparsity of the output matrix (k nnz coefficients per column).
    spcol(shape,k,'normalized', bool,'pos', bool): the optional parameters are set. By default both normalized and pos are false.

Parameters

shape	vector of size 2, to define the size of the input matrix.
k	the sparsity parameter.
'normalized',true	normalizes the projection image according to its Frobenius norm.
'normalized',false	(the default) no normalization.
'pos',true	skips the negative values (replaced by zero) of the input matrix.
'pos',false	(the default) negative values are not skipped.

Return values

spcol

projector.

Example

>> import matfaust.proj.spcol
>> rng(42)
>> M = rand(5, 5)
 
M =
 
    0.3745    0.1560    0.0206    0.1834    0.6119
    0.9507    0.0581    0.9699    0.3042    0.1395
    0.7320    0.8662    0.8324    0.5248    0.2921
    0.5987    0.6011    0.2123    0.4319    0.3664
    0.1560    0.7081    0.1818    0.2912    0.4561
 
>> p = spcol(size(M), 3, 'normalized', false);
>> p(M)
 
ans =
 
         0         0         0         0    0.6119
    0.9507         0    0.9699    0.3042         0
    0.7320    0.8662    0.8324    0.5248         0
    0.5987    0.6011    0.2123    0.4319    0.3664
         0    0.7081         0         0    0.4561
 
>>

Constructor & Destructor Documentation

spcol()

function matfaust::proj::spcol::spcol	(	shape	,
		k	,
		varargin
	)

---

The documentation for this class was generated from the following file:

• spcol.m