Skip to contents

Compute a (normalized) graph Laplacian from an adjacency matrix

Source: R/structure_descriptors.R
compute_graph_laplacian.Rd

Compute a (normalized) graph Laplacian from an adjacency matrix

Usage

compute_graph_laplacian(
  A,
  normalized = TRUE,
  symmetrize = TRUE,
  zero_diagonal = TRUE
)

Arguments

A

Square adjacency/affinity matrix (dense or sparse). Values should be nonnegative; the diagonal is set to 0 by default.

normalized

Logical; if TRUE (default), compute the normalized Laplacian \(L = I - D^{-1/2} A D^{-1/2}\). If FALSE, compute the unnormalized Laplacian \(L = D - A\).

symmetrize

Logical; if TRUE (default), symmetrize via `(A + t(A))/2`.

zero_diagonal

Logical; if TRUE (default), set `diag(A)=0` before forming the Laplacian.

Value

A sparse symmetric matrix (CsparseMatrix).