Kronecker Products on Sparse Matrices

Computes the generalised kronecker product of two arrays, X and Y.

kronecker.spam(X, Y, FUN = "*", make.dimnames = FALSE, ...)

Arguments

X

sparse matrix of class spam, a vector or a matrix.

Y

sparse matrix of class spam, a vector or a matrix.

FUN

a function; it may be a quoted string. See details

make.dimnames

Provide dimnames that are the product of the dimnames of X and Y.

...

optional arguments to be passed to FUN.

Value

An array A with dimensions

dim(X) * dim(Y).

Details

The sparsity structure is determined by the ordinary %x%. Hence, the result of kronecker(X, Y, FUN = "+") is different depending on the input.

Author

Reinhard Furrer

Examples

# Starting with non-spam objects, we get a spam matrix
kronecker.spam( diag(2), array(1:4, c(2, 2)))
#>      [,1] [,2] [,3] [,4]
#> [1,]    1    3    0    0
#> [2,]    2    4    0    0
#> [3,]    0    0    1    3
#> [4,]    0    0    2    4
#> Class 'spam' (32-bit)

kronecker( diag.spam(2), array(1:4, c(2, 2)))
#>      [,1] [,2] [,3] [,4]
#> [1,]    1    3    0    0
#> [2,]    2    4    0    0
#> [3,]    0    0    1    3
#> [4,]    0    0    2    4
#> Class 'spam' (32-bit)

# Notice the preservation of sparsity structure:
kronecker( diag.spam(2), array(1:4, c(2, 2)), FUN="+")
#> Warning: Sparseness structure of 'kronecker(X,Y)' preseved when applying 'FUN'.
#>      [,1] [,2] [,3] [,4]
#> [1,]    2    4    0    0
#> [2,]    3    5    0    0
#> [3,]    0    0    2    4
#> [4,]    0    0    3    5
#> Class 'spam' (32-bit)