center_graph: Center adjacency matrix

In dpmcsuss/iGraphMatch: Tools for Graph Matching

Center adjacency matrix

Description

Center the adjacency matrix by re-weighting edges according to a specified scheme

Usage

center_graph(A, scheme = c(-1, 1), use_splr = TRUE)

Arguments

A	A matrix, an igraph object. Adjacency matrix.
scheme	A character vector, number or pair of numbers. Default c(-1, 1). See Details.
use_splr	A boolean indicating whether to use the splrMatrix object when storing the centered graph. Defaults to TRUE.

Details

The options for scheme are

• "naive" Returns original A

• Integer: Returns A - A_{scheme} where A_{scheme} is the best rank-scheme approximation of A.

• A pair of scalars: Returns s * A + a such that the minimum of the returned matrix is min(scheme) and the maximum is max(scheme).

• "center": Same as scheme=c(-1,1)

Value

centered adjacency matrix as a splrMatrix if useSplr = TRUE, otherwise as a Matrix object.

Examples

A <- sample_correlated_gnp_pair(n = 10, corr = .5, p = .5)$graph1
center_graph(A, scheme = "naive")
center_graph(A, scheme = "center")
center_graph(A, scheme = 2)
center_graph(A, scheme = c(-4, 2))

dpmcsuss/iGraphMatch documentation built on May 22, 2024, 8:52 p.m.