Functions for dealing with sparse data matrices. More...

#include "gensvm_sparse.h"

Include dependency graph for gensvm_sparse.c:

Functions
struct GenSparse *	gensvm_init_sparse (void)
	Initialize a GenSparse structure. More...

void	gensvm_free_sparse (struct GenSparse *sp)
	Free an allocated GenSparse structure. More...

long	gensvm_count_nnz (double *A, long rows, long cols)
	Count the number of nonzeros in a matrix. More...

bool	gensvm_nnz_comparison (long nnz, long rows, long cols)
	Compare the number of nonzeros is such that sparsity if worth it. More...

bool	gensvm_could_sparse (double *A, long rows, long cols)
	Check if it is worthwile to convert to a sparse matrix. More...

struct GenSparse *	gensvm_dense_to_sparse (double *A, long rows, long cols)
	Convert a dense matrix to a GenSparse structure if advantageous. More...

double *	gensvm_sparse_to_dense (struct GenSparse *A)
	Convert a GenSparse structure to a dense matrix. More...

Detailed Description

Functions for dealing with sparse data matrices.

Author: G.J.J. van den Burg

Date: 2016-10-11

Copyright: Copyright 2016, G.J.J. van den Burg.

This file is part of GenSVM.

GenSVM is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

GenSVM is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with GenSVM. If not, see http://www.gnu.org/licenses/.

Definition in file gensvm_sparse.c.

Function Documentation

◆ gensvm_could_sparse()

bool gensvm_could_sparse	(	double *	A,
		long	rows,
		long	cols
	)

Check if it is worthwile to convert to a sparse matrix.

It is only worth to convert to a sparse matrix if the amount of sparsity is sufficient. For this to be the case, the number of nonzeros must be smaller than $(n_{row}(n_{col} - 1) - 1)/2$ . This is tested here. If the amount of nonzero entries is small enough, the function returns the number of nonzeros. If it is too big, it returns -1.

See also: gensvm_nnz_comparison()

Parameters

[in]	A	matrix in dense format (RowMajor order)
[in]	rows	number of rows of A
[in]	cols	number of columns of A

Returns: whether or not sparsity is worth it

Definition at line 129 of file gensvm_sparse.c.

Here is the call graph for this function:

◆ gensvm_count_nnz()

long gensvm_count_nnz	(	double *	A,
		long	rows,
		long	cols
	)

Count the number of nonzeros in a matrix.

This is a utility function to count the number of nonzeros in a dense matrix. This is simply done by comparing with 0.0.

Parameters

[in]	A	a dense matrix (RowMajor order)
[in]	rows	the number of rows of A
[in]	cols	the number of columns of A

Returns: the number of nonzeros in A

Definition at line 84 of file gensvm_sparse.c.

◆ gensvm_dense_to_sparse()

struct GenSparse* gensvm_dense_to_sparse	(	double *	A,
		long	rows,
		long	cols
	)

Convert a dense matrix to a GenSparse structure if advantageous.

This utility function can be used to convert a dense matrix to a sparse matrix in the form of a GenSparse struture. Note that the allocated memory must be freed by the caller. The user should first check whether using a sparse matrix is worth it by calling gensvm_could_sparse().

Parameters

[in]	A	a dense matrix in RowMajor order
[in]	rows	number of rows of the matrix A
[in]	cols	number of columns of the matrix A

Returns: a GenSparse struct

Definition at line 150 of file gensvm_sparse.c.

Here is the call graph for this function:

◆ gensvm_free_sparse()

void gensvm_free_sparse ( struct GenSparse * sp )

Free an allocated GenSparse structure.

Simply free a previously allocated GenSparse structure by freeing all of its components. Finally, the structure itself is freed, and the pointer is set to NULL for safety.

Parameters

[in] sp GenSparse structure to free

Definition at line 62 of file gensvm_sparse.c.

◆ gensvm_init_sparse()

struct GenSparse* gensvm_init_sparse ( void )

Initialize a GenSparse structure.

A GenSparse structure is used to hold a sparse data matrix. We work with Compressed Row Storage (CSR) storage, also known as old Yale format.

Returns: initialized GenSparse instance

Definition at line 38 of file gensvm_sparse.c.

◆ gensvm_nnz_comparison()

bool gensvm_nnz_comparison	(	long	nnz,
		long	rows,
		long	cols
	)

Compare the number of nonzeros is such that sparsity if worth it.

This is a utility function, see gensvm_could_sparse() for more info.

Parameters

[in]	nnz	number of nonzero elements
[in]	rows	number of rows
[in]	cols	number of columns

Returns: whether or not sparsity is worth it

Definition at line 105 of file gensvm_sparse.c.

◆ gensvm_sparse_to_dense()

double* gensvm_sparse_to_dense ( struct GenSparse * A )

Convert a GenSparse structure to a dense matrix.

This function converts a GenSparse structure back to a normal dense matrix in RowMajor order. Note that the allocated memory must be freed by the caller.

Parameters

[in] A a GenSparse structure

Returns: a dense matrix

Definition at line 199 of file gensvm_sparse.c.