Skip to contents

Symmetric Eigendecomposition

Source: R/api.R
nv_eigh.Rd

Computes the eigendecomposition of a symmetric matrix operand of shape (n, n): $$A = \mathrm{vectors} \, \mathrm{diag}(\mathrm{values}) \, \mathrm{vectors}^\top.$$ Only the lower triangle of operand is read. The columns of vectors are the (orthonormal) eigenvectors and values is the length-n vector of (real) eigenvalues in ascending order. Output names and order match base::eigen().

Usage

nv_eigh(operand)

Arguments

operand

(arrayish)
Symmetric square matrix of floating-point data type.

Value

Named list with elements values (length n) and vectors (shape (n, n)). Both have the same dtype as the input.

See also

prim_eigh(), base::eigen()

Examples

x <- nv_matrix(c(2, 1, 1, 2), nrow = 2, dtype = "f64")
nv_eigh(x)
#> $values
#> AnvlArray
#>  1
#>  3
#> [ CPUf64{2} ] 
#> 
#> $vectors
#> AnvlArray
#>  -0.7071  0.7071
#>   0.7071  0.7071
#> [ CPUf64{2,2} ] 
#>