Matrix operations

Runs any of TDA's matrix commands on R matrices and returns R matrices. This is a generic interface rather than one function per operation: R already has %*%, solve and eigen, so named wrappers would add nothing, but the whole of TDA's matrix language stays reachable.

Usage

tda_mat(op, ..., out = "R", options = list(), dir = tempfile("tda"))

Arguments

op

the command name. TDA's matrix language has about eighty: mmul multiplies, minvs and mginv invert, mtransp transposes, mev and mevs give eigenvalues, msvd the singular value decomposition, mchol a Cholesky factor, mrank the rank. Note the names: inversion is minvs, not minv, and transposition mtransp, not mtra.

...

input matrices, in the order the command expects.

out

names of the result matrices to read back.

options

a named list of further TDA options, passed through.

dir

working directory.

Value

A matrix if one result was asked for, otherwise a named list of them.

How it works

The inputs are written as TDA matrices named A, B, ... in order, the command is run, and every matrix named in out is read back. TDA's convention is that results come last in the argument list, so mmul(A,B,R) multiplies A by B into R.

tda_mat("mmul", A, B, out = "R")            # A 
tda_mat("mev", A, out = c("ER", "EI"))      # eigenvalues, real and imaginary

See also

Other smoothing: tda_integrate(), tda_interp(), tda_isotonic(), tda_sma(), tda_smd(), tda_spl()

Examples

A <- matrix(1:4, 2)
B <- diag(2)
tda_mat("mmul", A, B, out = "R")
#>      [,1] [,2]
#> [1,]    1    3
#> [2,]    2    4